code stringlengths 87 55.2k | code_codestyle int64 0 349 | style_context stringlengths 135 49.1k | style_context_codestyle int64 0 349 | label int64 0 1 |
|---|---|---|---|---|
"""simple docstring"""
import argparse
from diffusers.pipelines.stable_diffusion.convert_from_ckpt import download_controlnet_from_original_ckpt
if __name__ == "__main__":
__a = argparse.ArgumentParser()
parser.add_argument(
"--checkpoint_path", default=None, type=str, required=True, help="Path to the checkpoint to convert."
)
parser.add_argument(
"--original_config_file",
type=str,
required=True,
help="The YAML config file corresponding to the original architecture.",
)
parser.add_argument(
"--num_in_channels",
default=None,
type=int,
help="The number of input channels. If `None` number of input channels will be automatically inferred.",
)
parser.add_argument(
"--image_size",
default=5_12,
type=int,
help=(
"The image size that the model was trained on. Use 512 for Stable Diffusion v1.X and Stable Siffusion v2"
" Base. Use 768 for Stable Diffusion v2."
),
)
parser.add_argument(
"--extract_ema",
action="store_true",
help=(
"Only relevant for checkpoints that have both EMA and non-EMA weights. Whether to extract the EMA weights"
" or not. Defaults to `False`. Add `--extract_ema` to extract the EMA weights. EMA weights usually yield"
" higher quality images for inference. Non-EMA weights are usually better to continue fine-tuning."
),
)
parser.add_argument(
"--upcast_attention",
action="store_true",
help=(
"Whether the attention computation should always be upcasted. This is necessary when running stable"
" diffusion 2.1."
),
)
parser.add_argument(
"--from_safetensors",
action="store_true",
help="If `--checkpoint_path` is in `safetensors` format, load checkpoint with safetensors instead of PyTorch.",
)
parser.add_argument(
"--to_safetensors",
action="store_true",
help="Whether to store pipeline in safetensors format or not.",
)
parser.add_argument("--dump_path", default=None, type=str, required=True, help="Path to the output model.")
parser.add_argument("--device", type=str, help="Device to use (e.g. cpu, cuda:0, cuda:1, etc.)")
def A_ ( _lowercase ):
'''simple docstring'''
if string == "True":
return True
elif string == "False":
return False
else:
raise ValueError(f"""could not parse string as bool {string}""" )
parser.add_argument(
"--use_linear_projection", help="Override for use linear projection", required=False, type=parse_bool
)
parser.add_argument("--cross_attention_dim", help="Override for cross attention_dim", required=False, type=int)
__a = parser.parse_args()
__a = download_controlnet_from_original_ckpt(
checkpoint_path=args.checkpoint_path,
original_config_file=args.original_config_file,
image_size=args.image_size,
extract_ema=args.extract_ema,
num_in_channels=args.num_in_channels,
upcast_attention=args.upcast_attention,
from_safetensors=args.from_safetensors,
device=args.device,
use_linear_projection=args.use_linear_projection,
cross_attention_dim=args.cross_attention_dim,
)
controlnet.save_pretrained(args.dump_path, safe_serialization=args.to_safetensors)
| 66 |
"""simple docstring"""
import re
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Optional[int] = re.compile(
r"""^(?:0|94|\+94|0{2}94)""" r"""7(0|1|2|4|5|6|7|8)""" r"""(-| |)""" r"""\d{7}$""" )
return bool(re.search(_lowercase, _lowercase ) )
if __name__ == "__main__":
__a = "0094702343221"
print(is_sri_lankan_phone_number(phone))
| 66 | 1 |
"""simple docstring"""
from unittest.mock import Mock, patch
from file_transfer.send_file import send_file
@patch("""socket.socket""" )
@patch("""builtins.open""" )
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :str = Mock()
snake_case_ :Tuple = conn, Mock()
snake_case_ :Any = iter([1, None] )
snake_case_ :str = lambda _lowercase : next(_lowercase )
# ===== invoke =====
send_file(filename="""mytext.txt""", testing=_lowercase )
# ===== ensurance =====
sock.assert_called_once()
sock.return_value.bind.assert_called_once()
sock.return_value.listen.assert_called_once()
sock.return_value.accept.assert_called_once()
conn.recv.assert_called_once()
file.return_value.__enter__.assert_called_once()
file.return_value.__enter__.return_value.read.assert_called()
conn.send.assert_called_once()
conn.close.assert_called_once()
sock.return_value.shutdown.assert_called_once()
sock.return_value.close.assert_called_once()
| 66 |
"""simple docstring"""
import argparse
import json
import os
import pickle
import shutil
import numpy as np
import torch
from distiller import Distiller
from lm_seqs_dataset import LmSeqsDataset
from transformers import (
BertConfig,
BertForMaskedLM,
BertTokenizer,
DistilBertConfig,
DistilBertForMaskedLM,
DistilBertTokenizer,
GPTaConfig,
GPTaLMHeadModel,
GPTaTokenizer,
RobertaConfig,
RobertaForMaskedLM,
RobertaTokenizer,
)
from utils import git_log, init_gpu_params, logger, set_seed
__a = {
"distilbert": (DistilBertConfig, DistilBertForMaskedLM, DistilBertTokenizer),
"roberta": (RobertaConfig, RobertaForMaskedLM, RobertaTokenizer),
"bert": (BertConfig, BertForMaskedLM, BertTokenizer),
"gpt2": (GPTaConfig, GPTaLMHeadModel, GPTaTokenizer),
}
def A_ ( _lowercase ):
'''simple docstring'''
assert (args.mlm and args.alpha_mlm > 0.0) or (not args.mlm and args.alpha_mlm == 0.0)
assert (args.alpha_mlm > 0.0 and args.alpha_clm == 0.0) or (args.alpha_mlm == 0.0 and args.alpha_clm > 0.0)
if args.mlm:
assert os.path.isfile(args.token_counts )
assert (args.student_type in ["roberta", "distilbert"]) and (args.teacher_type in ["roberta", "bert"])
else:
assert (args.student_type in ["gpt2"]) and (args.teacher_type in ["gpt2"])
assert args.teacher_type == args.student_type or (
args.student_type == "distilbert" and args.teacher_type == "bert"
)
assert os.path.isfile(args.student_config )
if args.student_pretrained_weights is not None:
assert os.path.isfile(args.student_pretrained_weights )
if args.freeze_token_type_embds:
assert args.student_type in ["roberta"]
assert args.alpha_ce >= 0.0
assert args.alpha_mlm >= 0.0
assert args.alpha_clm >= 0.0
assert args.alpha_mse >= 0.0
assert args.alpha_cos >= 0.0
assert args.alpha_ce + args.alpha_mlm + args.alpha_clm + args.alpha_mse + args.alpha_cos > 0.0
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
if args.student_type == "roberta":
snake_case_ :Tuple = False
elif args.student_type == "gpt2":
snake_case_ :Union[str, Any] = False
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
if args.student_type == "roberta":
snake_case_ :List[str] = False
def A_ ( ):
'''simple docstring'''
snake_case_ :Union[str, Any] = argparse.ArgumentParser(description="""Training""" )
parser.add_argument("""--force""", action="""store_true""", help="""Overwrite dump_path if it already exists.""" )
parser.add_argument(
"""--dump_path""", type=_lowercase, required=_lowercase, help="""The output directory (log, checkpoints, parameters, etc.)""" )
parser.add_argument(
"""--data_file""", type=_lowercase, required=_lowercase, help="""The binarized file (tokenized + tokens_to_ids) and grouped by sequence.""", )
parser.add_argument(
"""--student_type""", type=_lowercase, choices=["""distilbert""", """roberta""", """gpt2"""], required=_lowercase, help="""The student type (DistilBERT, RoBERTa).""", )
parser.add_argument("""--student_config""", type=_lowercase, required=_lowercase, help="""Path to the student configuration.""" )
parser.add_argument(
"""--student_pretrained_weights""", default=_lowercase, type=_lowercase, help="""Load student initialization checkpoint.""" )
parser.add_argument(
"""--teacher_type""", choices=["""bert""", """roberta""", """gpt2"""], required=_lowercase, help="""Teacher type (BERT, RoBERTa).""" )
parser.add_argument("""--teacher_name""", type=_lowercase, required=_lowercase, help="""The teacher model.""" )
parser.add_argument("""--temperature""", default=2.0, type=_lowercase, help="""Temperature for the softmax temperature.""" )
parser.add_argument(
"""--alpha_ce""", default=0.5, type=_lowercase, help="""Linear weight for the distillation loss. Must be >=0.""" )
parser.add_argument(
"""--alpha_mlm""", default=0.0, type=_lowercase, help="""Linear weight for the MLM loss. Must be >=0. Should be used in conjunction with `mlm` flag.""", )
parser.add_argument("""--alpha_clm""", default=0.5, type=_lowercase, help="""Linear weight for the CLM loss. Must be >=0.""" )
parser.add_argument("""--alpha_mse""", default=0.0, type=_lowercase, help="""Linear weight of the MSE loss. Must be >=0.""" )
parser.add_argument(
"""--alpha_cos""", default=0.0, type=_lowercase, help="""Linear weight of the cosine embedding loss. Must be >=0.""" )
parser.add_argument(
"""--mlm""", action="""store_true""", help="""The LM step: MLM or CLM. If `mlm` is True, the MLM is used over CLM.""" )
parser.add_argument(
"""--mlm_mask_prop""", default=0.15, type=_lowercase, help="""Proportion of tokens for which we need to make a prediction.""", )
parser.add_argument("""--word_mask""", default=0.8, type=_lowercase, help="""Proportion of tokens to mask out.""" )
parser.add_argument("""--word_keep""", default=0.1, type=_lowercase, help="""Proportion of tokens to keep.""" )
parser.add_argument("""--word_rand""", default=0.1, type=_lowercase, help="""Proportion of tokens to randomly replace.""" )
parser.add_argument(
"""--mlm_smoothing""", default=0.7, type=_lowercase, help="""Smoothing parameter to emphasize more rare tokens (see XLM, similar to word2vec).""", )
parser.add_argument("""--token_counts""", type=_lowercase, help="""The token counts in the data_file for MLM.""" )
parser.add_argument(
"""--restrict_ce_to_mask""", action="""store_true""", help="""If true, compute the distillation loss only the [MLM] prediction distribution.""", )
parser.add_argument(
"""--freeze_pos_embs""", action="""store_true""", help="""Freeze positional embeddings during distillation. For student_type in ['roberta', 'gpt2'] only.""", )
parser.add_argument(
"""--freeze_token_type_embds""", action="""store_true""", help="""Freeze token type embeddings during distillation if existent. For student_type in ['roberta'] only.""", )
parser.add_argument("""--n_epoch""", type=_lowercase, default=3, help="""Number of pass on the whole dataset.""" )
parser.add_argument("""--batch_size""", type=_lowercase, default=5, help="""Batch size (for each process).""" )
parser.add_argument(
"""--group_by_size""", action="""store_false""", help="""If true, group sequences that have similar length into the same batch. Default is true.""", )
parser.add_argument(
"""--gradient_accumulation_steps""", type=_lowercase, default=50, help="""Gradient accumulation for larger training batches.""", )
parser.add_argument("""--warmup_prop""", default=0.05, type=_lowercase, help="""Linear warmup proportion.""" )
parser.add_argument("""--weight_decay""", default=0.0, type=_lowercase, help="""Weight decay if we apply some.""" )
parser.add_argument("""--learning_rate""", default=5e-4, type=_lowercase, help="""The initial learning rate for Adam.""" )
parser.add_argument("""--adam_epsilon""", default=1e-6, type=_lowercase, help="""Epsilon for Adam optimizer.""" )
parser.add_argument("""--max_grad_norm""", default=5.0, type=_lowercase, help="""Max gradient norm.""" )
parser.add_argument("""--initializer_range""", default=0.02, type=_lowercase, help="""Random initialization range.""" )
parser.add_argument(
"""--fp16""", action="""store_true""", help="""Whether to use 16-bit (mixed) precision (through NVIDIA apex) instead of 32-bit""", )
parser.add_argument(
"""--fp16_opt_level""", type=_lowercase, default="""O1""", help=(
"""For fp16: Apex AMP optimization level selected in ['O0', 'O1', 'O2', and 'O3']."""
"""See details at https://nvidia.github.io/apex/amp.html"""
), )
parser.add_argument("""--n_gpu""", type=_lowercase, default=1, help="""Number of GPUs in the node.""" )
parser.add_argument("""--local_rank""", type=_lowercase, default=-1, help="""Distributed training - Local rank""" )
parser.add_argument("""--seed""", type=_lowercase, default=56, help="""Random seed""" )
parser.add_argument("""--log_interval""", type=_lowercase, default=500, help="""Tensorboard logging interval.""" )
parser.add_argument("""--checkpoint_interval""", type=_lowercase, default=4000, help="""Checkpoint interval.""" )
snake_case_ :Tuple = parser.parse_args()
sanity_checks(_lowercase )
# ARGS #
init_gpu_params(_lowercase )
set_seed(_lowercase )
if args.is_master:
if os.path.exists(args.dump_path ):
if not args.force:
raise ValueError(
f"""Serialization dir {args.dump_path} already exists, but you have not precised wheter to overwrite"""
""" itUse `--force` if you want to overwrite it""" )
else:
shutil.rmtree(args.dump_path )
if not os.path.exists(args.dump_path ):
os.makedirs(args.dump_path )
logger.info(f"""Experiment will be dumped and logged in {args.dump_path}""" )
# SAVE PARAMS #
logger.info(f"""Param: {args}""" )
with open(os.path.join(args.dump_path, """parameters.json""" ), """w""" ) as f:
json.dump(vars(_lowercase ), _lowercase, indent=4 )
git_log(args.dump_path )
snake_case_, snake_case_, snake_case_ :Any = MODEL_CLASSES[args.student_type]
snake_case_, snake_case_, snake_case_ :int = MODEL_CLASSES[args.teacher_type]
# TOKENIZER #
snake_case_ :Any = teacher_tokenizer_class.from_pretrained(args.teacher_name )
snake_case_ :Optional[Any] = {}
for tok_name, tok_symbol in tokenizer.special_tokens_map.items():
snake_case_ :Union[str, Any] = tokenizer.all_special_tokens.index(_lowercase )
snake_case_ :Union[str, Any] = tokenizer.all_special_ids[idx]
logger.info(f"""Special tokens {special_tok_ids}""" )
snake_case_ :str = special_tok_ids
snake_case_ :Any = tokenizer.max_model_input_sizes[args.teacher_name]
# DATA LOADER #
logger.info(f"""Loading data from {args.data_file}""" )
with open(args.data_file, """rb""" ) as fp:
snake_case_ :str = pickle.load(_lowercase )
if args.mlm:
logger.info(f"""Loading token counts from {args.token_counts} (already pre-computed)""" )
with open(args.token_counts, """rb""" ) as fp:
snake_case_ :Optional[Any] = pickle.load(_lowercase )
snake_case_ :Tuple = np.maximum(_lowercase, 1 ) ** -args.mlm_smoothing
for idx in special_tok_ids.values():
snake_case_ :Optional[int] = 0.0 # do not predict special tokens
snake_case_ :int = torch.from_numpy(_lowercase )
else:
snake_case_ :List[str] = None
snake_case_ :Optional[int] = LmSeqsDataset(params=_lowercase, data=_lowercase )
logger.info("""Data loader created.""" )
# STUDENT #
logger.info(f"""Loading student config from {args.student_config}""" )
snake_case_ :List[Any] = student_config_class.from_pretrained(args.student_config )
snake_case_ :Union[str, Any] = True
if args.student_pretrained_weights is not None:
logger.info(f"""Loading pretrained weights from {args.student_pretrained_weights}""" )
snake_case_ :List[str] = student_model_class.from_pretrained(args.student_pretrained_weights, config=_lowercase )
else:
snake_case_ :Optional[int] = student_model_class(_lowercase )
if args.n_gpu > 0:
student.to(f"""cuda:{args.local_rank}""" )
logger.info("""Student loaded.""" )
# TEACHER #
snake_case_ :Dict = teacher_model_class.from_pretrained(args.teacher_name, output_hidden_states=_lowercase )
if args.n_gpu > 0:
teacher.to(f"""cuda:{args.local_rank}""" )
logger.info(f"""Teacher loaded from {args.teacher_name}.""" )
# FREEZING #
if args.freeze_pos_embs:
freeze_pos_embeddings(_lowercase, _lowercase )
if args.freeze_token_type_embds:
freeze_token_type_embeddings(_lowercase, _lowercase )
# SANITY CHECKS #
assert student.config.vocab_size == teacher.config.vocab_size
assert student.config.hidden_size == teacher.config.hidden_size
assert student.config.max_position_embeddings == teacher.config.max_position_embeddings
if args.mlm:
assert token_probs.size(0 ) == stu_architecture_config.vocab_size
# DISTILLER #
torch.cuda.empty_cache()
snake_case_ :Optional[int] = Distiller(
params=_lowercase, dataset=_lowercase, token_probs=_lowercase, student=_lowercase, teacher=_lowercase )
distiller.train()
logger.info("""Let's go get some drinks.""" )
if __name__ == "__main__":
main()
| 66 | 1 |
"""simple docstring"""
def A_ ( _lowercase ):
'''simple docstring'''
return sum(i for i in range(1, number // 2 + 1 ) if number % i == 0 ) == number
if __name__ == "__main__":
print("Program to check whether a number is a Perfect number or not...")
__a = int(input("Enter number: ").strip())
print(F"""{number} is {'' if perfect(number) else 'not '}a Perfect Number.""")
| 66 |
"""simple docstring"""
import json
import os
import subprocess
import unittest
from ast import literal_eval
import pytest
from parameterized import parameterized, parameterized_class
from . import is_sagemaker_available
if is_sagemaker_available():
from sagemaker import Session, TrainingJobAnalytics
from sagemaker.huggingface import HuggingFace
@pytest.mark.skipif(
literal_eval(os.getenv("""TEST_SAGEMAKER""" , """False""" ) ) is not True , reason="""Skipping test because should only be run when releasing minor transformers version""" , )
@pytest.mark.usefixtures("""sm_env""" )
@parameterized_class(
[
{
"""framework""": """pytorch""",
"""script""": """run_glue_model_parallelism.py""",
"""model_name_or_path""": """roberta-large""",
"""instance_type""": """ml.p3dn.24xlarge""",
"""results""": {"""train_runtime""": 1_6_0_0, """eval_accuracy""": 0.3, """eval_loss""": 1.2},
},
{
"""framework""": """pytorch""",
"""script""": """run_glue.py""",
"""model_name_or_path""": """roberta-large""",
"""instance_type""": """ml.p3dn.24xlarge""",
"""results""": {"""train_runtime""": 1_6_0_0, """eval_accuracy""": 0.3, """eval_loss""": 1.2},
},
] )
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: Any ) -> str:
if self.framework == "pytorch":
subprocess.run(
f"""cp ./examples/pytorch/text-classification/run_glue.py {self.env.test_path}/run_glue.py""".split() , encoding="""utf-8""" , check=snake_case , )
assert hasattr(self , """env""" )
def lowerCAmelCase_ ( self: int , snake_case: Dict ) -> List[Any]:
# configuration for running training on smdistributed Model Parallel
snake_case_ :Tuple = {
"""enabled""": True,
"""processes_per_host""": 8,
}
snake_case_ :List[Any] = {
"""enabled""": True,
"""parameters""": {
"""microbatches""": 4,
"""placement_strategy""": """spread""",
"""pipeline""": """interleaved""",
"""optimize""": """speed""",
"""partitions""": 4,
"""ddp""": True,
},
}
snake_case_ :Tuple = {"""smdistributed""": {"""modelparallel""": smp_options}, """mpi""": mpi_options}
snake_case_ :Any = """trainer""" if self.script == """run_glue.py""" else """smtrainer"""
# creates estimator
return HuggingFace(
entry_point=self.script , source_dir=self.env.test_path , role=self.env.role , image_uri=self.env.image_uri , base_job_name=f"""{self.env.base_job_name}-{instance_count}-smp-{name_extension}""" , instance_count=snake_case , instance_type=self.instance_type , debugger_hook_config=snake_case , hyperparameters={
**self.env.hyperparameters,
"""model_name_or_path""": self.model_name_or_path,
"""max_steps""": 500,
} , metric_definitions=self.env.metric_definitions , distribution=snake_case , py_version="""py36""" , )
def lowerCAmelCase_ ( self: Any , snake_case: Tuple ) -> List[str]:
TrainingJobAnalytics(snake_case ).export_csv(f"""{self.env.test_path}/{job_name}_metrics.csv""" )
@parameterized.expand([(1,)] )
def lowerCAmelCase_ ( self: Dict , snake_case: Dict ) -> List[Any]:
# create estimator
snake_case_ :List[Any] = self.create_estimator(snake_case )
# run training
estimator.fit()
# result dataframe
snake_case_ :Any = TrainingJobAnalytics(estimator.latest_training_job.name ).dataframe()
# extract kpis
snake_case_ :Tuple = list(result_metrics_df[result_metrics_df.metric_name == """eval_accuracy"""]["""value"""] )
snake_case_ :Dict = list(result_metrics_df[result_metrics_df.metric_name == """eval_loss"""]["""value"""] )
# get train time from SageMaker job, this includes starting, preprocessing, stopping
snake_case_ :int = (
Session().describe_training_job(estimator.latest_training_job.name ).get("""TrainingTimeInSeconds""" , 999_999 )
)
# assert kpis
assert train_runtime <= self.results["train_runtime"]
assert all(t >= self.results["""eval_accuracy"""] for t in eval_accuracy )
assert all(t <= self.results["""eval_loss"""] for t in eval_loss )
# dump tests result into json file to share in PR
with open(f"""{estimator.latest_training_job.name}.json""" , """w""" ) as outfile:
json.dump({"""train_time""": train_runtime, """eval_accuracy""": eval_accuracy, """eval_loss""": eval_loss} , snake_case )
| 66 | 1 |
"""simple docstring"""
import importlib
import json
import os
from collections import OrderedDict
from typing import Dict, Optional, Union
# Build the list of all feature extractors
from ...configuration_utils import PretrainedConfig
from ...dynamic_module_utils import get_class_from_dynamic_module, resolve_trust_remote_code
from ...feature_extraction_utils import FeatureExtractionMixin
from ...utils import CONFIG_NAME, FEATURE_EXTRACTOR_NAME, get_file_from_repo, logging
from .auto_factory import _LazyAutoMapping
from .configuration_auto import (
CONFIG_MAPPING_NAMES,
AutoConfig,
model_type_to_module_name,
replace_list_option_in_docstrings,
)
__a = logging.get_logger(__name__)
__a = OrderedDict(
[
("audio-spectrogram-transformer", "ASTFeatureExtractor"),
("beit", "BeitFeatureExtractor"),
("chinese_clip", "ChineseCLIPFeatureExtractor"),
("clap", "ClapFeatureExtractor"),
("clip", "CLIPFeatureExtractor"),
("clipseg", "ViTFeatureExtractor"),
("conditional_detr", "ConditionalDetrFeatureExtractor"),
("convnext", "ConvNextFeatureExtractor"),
("cvt", "ConvNextFeatureExtractor"),
("data2vec-audio", "Wav2Vec2FeatureExtractor"),
("data2vec-vision", "BeitFeatureExtractor"),
("deformable_detr", "DeformableDetrFeatureExtractor"),
("deit", "DeiTFeatureExtractor"),
("detr", "DetrFeatureExtractor"),
("dinat", "ViTFeatureExtractor"),
("donut-swin", "DonutFeatureExtractor"),
("dpt", "DPTFeatureExtractor"),
("encodec", "EncodecFeatureExtractor"),
("flava", "FlavaFeatureExtractor"),
("glpn", "GLPNFeatureExtractor"),
("groupvit", "CLIPFeatureExtractor"),
("hubert", "Wav2Vec2FeatureExtractor"),
("imagegpt", "ImageGPTFeatureExtractor"),
("layoutlmv2", "LayoutLMv2FeatureExtractor"),
("layoutlmv3", "LayoutLMv3FeatureExtractor"),
("levit", "LevitFeatureExtractor"),
("maskformer", "MaskFormerFeatureExtractor"),
("mctct", "MCTCTFeatureExtractor"),
("mobilenet_v1", "MobileNetV1FeatureExtractor"),
("mobilenet_v2", "MobileNetV2FeatureExtractor"),
("mobilevit", "MobileViTFeatureExtractor"),
("nat", "ViTFeatureExtractor"),
("owlvit", "OwlViTFeatureExtractor"),
("perceiver", "PerceiverFeatureExtractor"),
("poolformer", "PoolFormerFeatureExtractor"),
("regnet", "ConvNextFeatureExtractor"),
("resnet", "ConvNextFeatureExtractor"),
("segformer", "SegformerFeatureExtractor"),
("sew", "Wav2Vec2FeatureExtractor"),
("sew-d", "Wav2Vec2FeatureExtractor"),
("speech_to_text", "Speech2TextFeatureExtractor"),
("speecht5", "SpeechT5FeatureExtractor"),
("swiftformer", "ViTFeatureExtractor"),
("swin", "ViTFeatureExtractor"),
("swinv2", "ViTFeatureExtractor"),
("table-transformer", "DetrFeatureExtractor"),
("timesformer", "VideoMAEFeatureExtractor"),
("tvlt", "TvltFeatureExtractor"),
("unispeech", "Wav2Vec2FeatureExtractor"),
("unispeech-sat", "Wav2Vec2FeatureExtractor"),
("van", "ConvNextFeatureExtractor"),
("videomae", "VideoMAEFeatureExtractor"),
("vilt", "ViltFeatureExtractor"),
("vit", "ViTFeatureExtractor"),
("vit_mae", "ViTFeatureExtractor"),
("vit_msn", "ViTFeatureExtractor"),
("wav2vec2", "Wav2Vec2FeatureExtractor"),
("wav2vec2-conformer", "Wav2Vec2FeatureExtractor"),
("wavlm", "Wav2Vec2FeatureExtractor"),
("whisper", "WhisperFeatureExtractor"),
("xclip", "CLIPFeatureExtractor"),
("yolos", "YolosFeatureExtractor"),
]
)
__a = _LazyAutoMapping(CONFIG_MAPPING_NAMES, FEATURE_EXTRACTOR_MAPPING_NAMES)
def A_ ( _lowercase ):
'''simple docstring'''
for module_name, extractors in FEATURE_EXTRACTOR_MAPPING_NAMES.items():
if class_name in extractors:
snake_case_ :Optional[int] = model_type_to_module_name(_lowercase )
snake_case_ :Tuple = importlib.import_module(f""".{module_name}""", """transformers.models""" )
try:
return getattr(_lowercase, _lowercase )
except AttributeError:
continue
for _, extractor in FEATURE_EXTRACTOR_MAPPING._extra_content.items():
if getattr(_lowercase, """__name__""", _lowercase ) == class_name:
return extractor
# We did not fine the class, but maybe it's because a dep is missing. In that case, the class will be in the main
# init and we return the proper dummy to get an appropriate error message.
snake_case_ :Optional[int] = importlib.import_module("""transformers""" )
if hasattr(_lowercase, _lowercase ):
return getattr(_lowercase, _lowercase )
return None
def A_ ( _lowercase, _lowercase = None, _lowercase = False, _lowercase = False, _lowercase = None, _lowercase = None, _lowercase = None, _lowercase = False, **_lowercase, ):
'''simple docstring'''
snake_case_ :Union[str, Any] = get_file_from_repo(
_lowercase, _lowercase, cache_dir=_lowercase, force_download=_lowercase, resume_download=_lowercase, proxies=_lowercase, use_auth_token=_lowercase, revision=_lowercase, local_files_only=_lowercase, )
if resolved_config_file is None:
logger.info(
"""Could not locate the feature extractor configuration file, will try to use the model config instead.""" )
return {}
with open(_lowercase, encoding="""utf-8""" ) as reader:
return json.load(_lowercase )
class lowerCamelCase :
'''simple docstring'''
def __init__( self: List[str] ) -> Tuple:
raise EnvironmentError(
"""AutoFeatureExtractor is designed to be instantiated """
"""using the `AutoFeatureExtractor.from_pretrained(pretrained_model_name_or_path)` method.""" )
@classmethod
@replace_list_option_in_docstrings(snake_case )
def lowerCAmelCase_ ( cls: List[Any] , snake_case: List[Any] , **snake_case: Any ) -> List[Any]:
snake_case_ :Optional[int] = kwargs.pop("""config""" , snake_case )
snake_case_ :Optional[int] = kwargs.pop("""trust_remote_code""" , snake_case )
snake_case_ :str = True
snake_case_, snake_case_ :int = FeatureExtractionMixin.get_feature_extractor_dict(snake_case , **snake_case )
snake_case_ :List[str] = config_dict.get("""feature_extractor_type""" , snake_case )
snake_case_ :List[Any] = None
if "AutoFeatureExtractor" in config_dict.get("""auto_map""" , {} ):
snake_case_ :str = config_dict["""auto_map"""]["""AutoFeatureExtractor"""]
# If we don't find the feature extractor class in the feature extractor config, let's try the model config.
if feature_extractor_class is None and feature_extractor_auto_map is None:
if not isinstance(snake_case , snake_case ):
snake_case_ :Union[str, Any] = AutoConfig.from_pretrained(snake_case , **snake_case )
# It could be in `config.feature_extractor_type``
snake_case_ :Dict = getattr(snake_case , """feature_extractor_type""" , snake_case )
if hasattr(snake_case , """auto_map""" ) and "AutoFeatureExtractor" in config.auto_map:
snake_case_ :str = config.auto_map["""AutoFeatureExtractor"""]
if feature_extractor_class is not None:
snake_case_ :str = feature_extractor_class_from_name(snake_case )
snake_case_ :Any = feature_extractor_auto_map is not None
snake_case_ :int = feature_extractor_class is not None or type(snake_case ) in FEATURE_EXTRACTOR_MAPPING
snake_case_ :Optional[int] = resolve_trust_remote_code(
snake_case , snake_case , snake_case , snake_case )
if has_remote_code and trust_remote_code:
snake_case_ :str = get_class_from_dynamic_module(
snake_case , snake_case , **snake_case )
snake_case_ :Union[str, Any] = kwargs.pop("""code_revision""" , snake_case )
if os.path.isdir(snake_case ):
feature_extractor_class.register_for_auto_class()
return feature_extractor_class.from_dict(snake_case , **snake_case )
elif feature_extractor_class is not None:
return feature_extractor_class.from_dict(snake_case , **snake_case )
# Last try: we use the FEATURE_EXTRACTOR_MAPPING.
elif type(snake_case ) in FEATURE_EXTRACTOR_MAPPING:
snake_case_ :Optional[int] = FEATURE_EXTRACTOR_MAPPING[type(snake_case )]
return feature_extractor_class.from_dict(snake_case , **snake_case )
raise ValueError(
f"""Unrecognized feature extractor in {pretrained_model_name_or_path}. Should have a """
f"""`feature_extractor_type` key in its {FEATURE_EXTRACTOR_NAME} of {CONFIG_NAME}, or one of the following """
f"""`model_type` keys in its {CONFIG_NAME}: {', '.join(c for c in FEATURE_EXTRACTOR_MAPPING_NAMES.keys() )}""" )
@staticmethod
def lowerCAmelCase_ ( snake_case: List[Any] , snake_case: int ) -> Optional[Any]:
FEATURE_EXTRACTOR_MAPPING.register(snake_case , snake_case )
| 66 |
"""simple docstring"""
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 lowerCamelCase :
'''simple docstring'''
def __init__( self: Dict , snake_case: Optional[Any] , snake_case: Tuple=13 , snake_case: Any=32 , snake_case: Union[str, Any]=2 , snake_case: Tuple=3 , snake_case: Union[str, Any]=16 , snake_case: Union[str, Any]=[1, 2, 1] , snake_case: Optional[Any]=[2, 2, 4] , snake_case: str=2 , snake_case: List[str]=2.0 , snake_case: Optional[int]=True , snake_case: Union[str, Any]=0.0 , snake_case: Optional[int]=0.0 , snake_case: Optional[Any]=0.1 , snake_case: List[str]="gelu" , snake_case: Any=False , snake_case: Optional[Any]=True , snake_case: Optional[int]=0.0_2 , snake_case: Any=1E-5 , snake_case: Optional[int]=True , snake_case: int=None , snake_case: Any=True , snake_case: str=10 , snake_case: Optional[Any]=8 , snake_case: Union[str, Any]=["stage1", "stage2", "stage3"] , snake_case: Tuple=[1, 2, 3] , ) -> Dict:
snake_case_ :Dict = parent
snake_case_ :List[Any] = batch_size
snake_case_ :Dict = image_size
snake_case_ :Dict = patch_size
snake_case_ :Tuple = num_channels
snake_case_ :List[Any] = embed_dim
snake_case_ :List[str] = depths
snake_case_ :str = num_heads
snake_case_ :Tuple = window_size
snake_case_ :Tuple = mlp_ratio
snake_case_ :int = qkv_bias
snake_case_ :Tuple = hidden_dropout_prob
snake_case_ :Optional[Any] = attention_probs_dropout_prob
snake_case_ :Dict = drop_path_rate
snake_case_ :Any = hidden_act
snake_case_ :Any = use_absolute_embeddings
snake_case_ :int = patch_norm
snake_case_ :List[Any] = layer_norm_eps
snake_case_ :Tuple = initializer_range
snake_case_ :str = is_training
snake_case_ :int = scope
snake_case_ :Tuple = use_labels
snake_case_ :Tuple = type_sequence_label_size
snake_case_ :str = encoder_stride
snake_case_ :List[Any] = out_features
snake_case_ :str = out_indices
def lowerCAmelCase_ ( self: Tuple ) -> Dict:
snake_case_ :Optional[int] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case_ :str = None
if self.use_labels:
snake_case_ :Optional[int] = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ :Union[str, Any] = self.get_config()
return config, pixel_values, labels
def lowerCAmelCase_ ( self: int ) -> Optional[Any]:
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 lowerCAmelCase_ ( self: List[Any] , snake_case: str , snake_case: int , snake_case: List[str] ) -> Any:
snake_case_ :Dict = MaskFormerSwinModel(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Tuple = model(snake_case )
snake_case_ :Dict = ((config.image_size // config.patch_size) ** 2) // (4 ** (len(config.depths ) - 1))
snake_case_ :Any = 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 lowerCAmelCase_ ( self: Optional[Any] , snake_case: int , snake_case: List[str] , snake_case: Tuple ) -> Union[str, Any]:
snake_case_ :Any = MaskFormerSwinBackbone(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Optional[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_ :Optional[Any] = ["""stem"""]
snake_case_ :str = MaskFormerSwinBackbone(config=snake_case )
def lowerCAmelCase_ ( self: List[str] ) -> Optional[Any]:
snake_case_ :Optional[int] = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_ :str = config_and_inputs
snake_case_ :Tuple = {"""pixel_values""": pixel_values}
return config, inputs_dict
@require_torch
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : Union[str, Any] = (
(
MaskFormerSwinModel,
MaskFormerSwinBackbone,
)
if is_torch_available()
else ()
)
_A : str = {"""feature-extraction""": MaskFormerSwinModel} if is_torch_available() else {}
_A : List[str] = False
_A : Any = False
_A : Dict = False
_A : List[Any] = False
_A : Optional[int] = False
def lowerCAmelCase_ ( self: Dict ) -> Any:
snake_case_ :str = MaskFormerSwinModelTester(self )
snake_case_ :Optional[Any] = 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 lowerCAmelCase_ ( self: List[str] ) -> Optional[int]:
pass
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Dict:
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 lowerCAmelCase_ ( self: Any ) -> Tuple:
return
def lowerCAmelCase_ ( self: Any ) -> Any:
snake_case_ :List[Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> int:
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 lowerCAmelCase_ ( self: str ) -> List[str]:
pass
@unittest.skip("""Swin does not support feedforward chunking""" )
def lowerCAmelCase_ ( self: int ) -> Optional[int]:
pass
def lowerCAmelCase_ ( self: List[str] ) -> List[Any]:
snake_case_, snake_case_ :List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :str = model_class(snake_case )
self.assertIsInstance(model.get_input_embeddings() , (nn.Module) )
snake_case_ :Dict = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(snake_case , nn.Linear ) )
def lowerCAmelCase_ ( self: Tuple ) -> Dict:
snake_case_, snake_case_ :int = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :Optional[int] = model_class(snake_case )
snake_case_ :str = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case_ :str = [*signature.parameters.keys()]
snake_case_ :str = ["""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 lowerCAmelCase_ ( self: List[Any] ) -> List[Any]:
pass
@unittest.skip(reason="""MaskFormerSwin is only used as an internal backbone""" )
def lowerCAmelCase_ ( self: Dict ) -> List[Any]:
pass
def lowerCAmelCase_ ( self: Union[str, Any] , snake_case: Union[str, Any] , snake_case: int , snake_case: Any , snake_case: List[str] ) -> str:
snake_case_ :List[str] = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :List[Any] = model(**self._prepare_for_class(snake_case , snake_case ) )
snake_case_ :Any = outputs.hidden_states
snake_case_ :Optional[int] = 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_ :str = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :int = (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 lowerCAmelCase_ ( self: List[Any] ) -> Optional[int]:
snake_case_, snake_case_ :Any = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :List[Any] = (
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_ :Tuple = 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_ :List[Any] = True
self.check_hidden_states_output(snake_case , snake_case , snake_case , snake_case )
def lowerCAmelCase_ ( self: Optional[Any] ) -> Tuple:
snake_case_, snake_case_ :int = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :List[Any] = 3
snake_case_ :List[Any] = (
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_ :Any = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :Tuple = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
snake_case_ :List[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_ :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 lowerCAmelCase_ ( self: Union[str, Any] ) -> List[str]:
pass
@unittest.skip(reason="""This will be fixed once MaskFormerSwin is replaced by native Swin""" )
def lowerCAmelCase_ ( self: List[str] ) -> str:
pass
@unittest.skip(reason="""This will be fixed once MaskFormerSwin is replaced by native Swin""" )
def lowerCAmelCase_ ( self: str ) -> List[Any]:
pass
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Optional[Any]:
snake_case_, snake_case_ :Dict = self.model_tester.prepare_config_and_inputs_for_common()
def set_nan_tensor_to_zero(snake_case: str ):
snake_case_ :Optional[int] = 0
return t
def check_equivalence(snake_case: List[Any] , snake_case: Union[str, Any] , snake_case: int , snake_case: Tuple={} ):
with torch.no_grad():
snake_case_ :List[Any] = model(**snake_case , return_dict=snake_case , **snake_case )
snake_case_ :Any = model(**snake_case , return_dict=snake_case , **snake_case ).to_tuple()
def recursive_check(snake_case: List[Any] , snake_case: int ):
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_ :int = model_class(snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Any = 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 )
snake_case_ :Tuple = self._prepare_for_class(snake_case , snake_case , return_labels=snake_case )
snake_case_ :Dict = self._prepare_for_class(snake_case , snake_case , return_labels=snake_case )
check_equivalence(snake_case , snake_case , snake_case )
snake_case_ :Tuple = self._prepare_for_class(snake_case , snake_case )
snake_case_ :Any = self._prepare_for_class(snake_case , snake_case )
check_equivalence(snake_case , snake_case , snake_case , {"""output_hidden_states""": True} )
snake_case_ :Dict = 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 lowerCamelCase ( unittest.TestCase , _lowerCAmelCase ):
'''simple docstring'''
_A : int = (MaskFormerSwinBackbone,) if is_torch_available() else ()
_A : Tuple = MaskFormerSwinConfig
def lowerCAmelCase_ ( self: List[str] ) -> Optional[int]:
snake_case_ :Optional[Any] = MaskFormerSwinModelTester(self )
def lowerCAmelCase_ ( self: int ) -> Optional[int]:
snake_case_, snake_case_ :Any = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Tuple = inputs_dict["""pixel_values"""].shape[0]
for backbone_class in self.all_model_classes:
snake_case_ :List[str] = backbone_class(snake_case )
backbone.to(snake_case )
backbone.eval()
snake_case_ :List[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_ :Union[str, Any] = 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_ :List[Any] = 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_ :List[Any] = backbone(**snake_case , output_attentions=snake_case )
self.assertIsNotNone(outputs.attentions )
| 66 | 1 |
"""simple docstring"""
from math import factorial
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
if n < k or k < 0:
raise ValueError("""Please enter positive integers for n and k where n >= k""" )
return factorial(_lowercase ) // (factorial(_lowercase ) * factorial(n - k ))
if __name__ == "__main__":
print(
"The number of five-card hands possible from a standard",
F"""fifty-two card deck is: {combinations(52, 5)}\n""",
)
print(
"If a class of 40 students must be arranged into groups of",
F"""4 for group projects, there are {combinations(40, 4)} ways""",
"to arrange them.\n",
)
print(
"If 10 teams are competing in a Formula One race, there",
F"""are {combinations(10, 3)} ways that first, second and""",
"third place can be awarded.",
)
| 66 |
"""simple docstring"""
import gc
import math
import unittest
import torch
from diffusers import UNetaDModel
from diffusers.utils import floats_tensor, logging, slow, torch_all_close, torch_device
from diffusers.utils.testing_utils import enable_full_determinism
from .test_modeling_common import ModelTesterMixin, UNetTesterMixin
__a = logging.get_logger(__name__)
enable_full_determinism()
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[Any] = UNetaDModel
_A : Union[str, Any] = """sample"""
@property
def lowerCAmelCase_ ( self: str ) -> Tuple:
snake_case_ :List[str] = 4
snake_case_ :Tuple = 3
snake_case_ :Optional[Any] = (32, 32)
snake_case_ :str = floats_tensor((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :Union[str, Any] = torch.tensor([10] ).to(snake_case )
return {"sample": noise, "timestep": time_step}
@property
def lowerCAmelCase_ ( self: List[str] ) -> Dict:
return (3, 32, 32)
@property
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
return (3, 32, 32)
def lowerCAmelCase_ ( self: Optional[int] ) -> Dict:
snake_case_ :Any = {
"""block_out_channels""": (32, 64),
"""down_block_types""": ("""DownBlock2D""", """AttnDownBlock2D"""),
"""up_block_types""": ("""AttnUpBlock2D""", """UpBlock2D"""),
"""attention_head_dim""": 3,
"""out_channels""": 3,
"""in_channels""": 3,
"""layers_per_block""": 2,
"""sample_size""": 32,
}
snake_case_ :Tuple = self.dummy_input
return init_dict, inputs_dict
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[str] = UNetaDModel
_A : Union[str, Any] = """sample"""
@property
def lowerCAmelCase_ ( self: str ) -> str:
snake_case_ :List[str] = 4
snake_case_ :Optional[int] = 4
snake_case_ :int = (32, 32)
snake_case_ :Any = floats_tensor((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :List[Any] = torch.tensor([10] ).to(snake_case )
return {"sample": noise, "timestep": time_step}
@property
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
return (4, 32, 32)
@property
def lowerCAmelCase_ ( self: List[Any] ) -> int:
return (4, 32, 32)
def lowerCAmelCase_ ( self: Union[str, Any] ) -> List[Any]:
snake_case_ :Dict = {
"""sample_size""": 32,
"""in_channels""": 4,
"""out_channels""": 4,
"""layers_per_block""": 2,
"""block_out_channels""": (32, 64),
"""attention_head_dim""": 32,
"""down_block_types""": ("""DownBlock2D""", """DownBlock2D"""),
"""up_block_types""": ("""UpBlock2D""", """UpBlock2D"""),
}
snake_case_ :List[str] = self.dummy_input
return init_dict, inputs_dict
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[Any]:
snake_case_, snake_case_ :List[str] = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case )
self.assertIsNotNone(snake_case )
self.assertEqual(len(loading_info["""missing_keys"""] ) , 0 )
model.to(snake_case )
snake_case_ :List[str] = model(**self.dummy_input ).sample
assert image is not None, "Make sure output is not None"
@unittest.skipIf(torch_device != """cuda""" , """This test is supposed to run on GPU""" )
def lowerCAmelCase_ ( self: Tuple ) -> Dict:
snake_case_, snake_case_ :Union[str, Any] = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case )
model.to(snake_case )
snake_case_ :Union[str, Any] = model(**self.dummy_input ).sample
assert image is not None, "Make sure output is not None"
@unittest.skipIf(torch_device != """cuda""" , """This test is supposed to run on GPU""" )
def lowerCAmelCase_ ( self: str ) -> Any:
# by defautl model loading will use accelerate as `low_cpu_mem_usage=True`
snake_case_, snake_case_ :List[str] = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case )
model_accelerate.to(snake_case )
model_accelerate.eval()
snake_case_ :List[Any] = torch.randn(
1 , model_accelerate.config.in_channels , model_accelerate.config.sample_size , model_accelerate.config.sample_size , generator=torch.manual_seed(0 ) , )
snake_case_ :int = noise.to(snake_case )
snake_case_ :str = torch.tensor([10] * noise.shape[0] ).to(snake_case )
snake_case_ :Optional[int] = model_accelerate(snake_case , snake_case )["""sample"""]
# two models don't need to stay in the device at the same time
del model_accelerate
torch.cuda.empty_cache()
gc.collect()
snake_case_, snake_case_ :str = UNetaDModel.from_pretrained(
"""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case , low_cpu_mem_usage=snake_case )
model_normal_load.to(snake_case )
model_normal_load.eval()
snake_case_ :int = model_normal_load(snake_case , snake_case )["""sample"""]
assert torch_all_close(snake_case , snake_case , rtol=1E-3 )
def lowerCAmelCase_ ( self: Tuple ) -> Any:
snake_case_ :Tuple = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" )
model.eval()
model.to(snake_case )
snake_case_ :Optional[int] = torch.randn(
1 , model.config.in_channels , model.config.sample_size , model.config.sample_size , generator=torch.manual_seed(0 ) , )
snake_case_ :int = noise.to(snake_case )
snake_case_ :List[Any] = torch.tensor([10] * noise.shape[0] ).to(snake_case )
with torch.no_grad():
snake_case_ :Union[str, Any] = model(snake_case , snake_case ).sample
snake_case_ :Optional[int] = output[0, -1, -3:, -3:].flatten().cpu()
# fmt: off
snake_case_ :Dict = torch.tensor([-1_3.3_2_5_8, -2_0.1_1_0_0, -1_5.9_8_7_3, -1_7.6_6_1_7, -2_3.0_5_9_6, -1_7.9_4_1_9, -1_3.3_6_7_5, -1_6.1_8_8_9, -1_2.3_8_0_0] )
# fmt: on
self.assertTrue(torch_all_close(snake_case , snake_case , rtol=1E-3 ) )
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[Any] = UNetaDModel
_A : List[Any] = """sample"""
@property
def lowerCAmelCase_ ( self: Union[str, Any] , snake_case: int=(32, 32) ) -> Tuple:
snake_case_ :Union[str, Any] = 4
snake_case_ :Any = 3
snake_case_ :int = floats_tensor((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :Any = torch.tensor(batch_size * [10] ).to(dtype=torch.intaa , device=snake_case )
return {"sample": noise, "timestep": time_step}
@property
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Any:
return (3, 32, 32)
@property
def lowerCAmelCase_ ( self: int ) -> Tuple:
return (3, 32, 32)
def lowerCAmelCase_ ( self: List[str] ) -> Tuple:
snake_case_ :List[Any] = {
"""block_out_channels""": [32, 64, 64, 64],
"""in_channels""": 3,
"""layers_per_block""": 1,
"""out_channels""": 3,
"""time_embedding_type""": """fourier""",
"""norm_eps""": 1E-6,
"""mid_block_scale_factor""": math.sqrt(2.0 ),
"""norm_num_groups""": None,
"""down_block_types""": [
"""SkipDownBlock2D""",
"""AttnSkipDownBlock2D""",
"""SkipDownBlock2D""",
"""SkipDownBlock2D""",
],
"""up_block_types""": [
"""SkipUpBlock2D""",
"""SkipUpBlock2D""",
"""AttnSkipUpBlock2D""",
"""SkipUpBlock2D""",
],
}
snake_case_ :int = self.dummy_input
return init_dict, inputs_dict
@slow
def lowerCAmelCase_ ( self: Optional[Any] ) -> List[Any]:
snake_case_, snake_case_ :List[Any] = UNetaDModel.from_pretrained("""google/ncsnpp-celebahq-256""" , output_loading_info=snake_case )
self.assertIsNotNone(snake_case )
self.assertEqual(len(loading_info["""missing_keys"""] ) , 0 )
model.to(snake_case )
snake_case_ :Any = self.dummy_input
snake_case_ :int = floats_tensor((4, 3) + (256, 256) ).to(snake_case )
snake_case_ :int = noise
snake_case_ :int = model(**snake_case )
assert image is not None, "Make sure output is not None"
@slow
def lowerCAmelCase_ ( self: str ) -> Dict:
snake_case_ :Dict = UNetaDModel.from_pretrained("""google/ncsnpp-celebahq-256""" )
model.to(snake_case )
snake_case_ :List[str] = 4
snake_case_ :Optional[int] = 3
snake_case_ :List[str] = (256, 256)
snake_case_ :Tuple = torch.ones((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :str = torch.tensor(batch_size * [1E-4] ).to(snake_case )
with torch.no_grad():
snake_case_ :Dict = model(snake_case , snake_case ).sample
snake_case_ :int = output[0, -3:, -3:, -1].flatten().cpu()
# fmt: off
snake_case_ :Optional[Any] = torch.tensor([-4_8_4_2.8_6_9_1, -6_4_9_9.6_6_3_1, -3_8_0_0.1_9_5_3, -7_9_7_8.2_6_8_6, -1_0_9_8_0.7_1_2_9, -2_0_0_2_8.8_5_3_5, 8_1_4_8.2_8_2_2, 2_3_4_2.2_9_0_5, 5_6_7.7_6_0_8] )
# fmt: on
self.assertTrue(torch_all_close(snake_case , snake_case , rtol=1E-2 ) )
def lowerCAmelCase_ ( self: List[str] ) -> List[Any]:
snake_case_ :Optional[Any] = UNetaDModel.from_pretrained("""fusing/ncsnpp-ffhq-ve-dummy-update""" )
model.to(snake_case )
snake_case_ :Optional[int] = 4
snake_case_ :Optional[Any] = 3
snake_case_ :Optional[Any] = (32, 32)
snake_case_ :Dict = torch.ones((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :Any = torch.tensor(batch_size * [1E-4] ).to(snake_case )
with torch.no_grad():
snake_case_ :str = model(snake_case , snake_case ).sample
snake_case_ :int = output[0, -3:, -3:, -1].flatten().cpu()
# fmt: off
snake_case_ :int = torch.tensor([-0.0_3_2_5, -0.0_9_0_0, -0.0_8_6_9, -0.0_3_3_2, -0.0_7_2_5, -0.0_2_7_0, -0.0_1_0_1, 0.0_2_2_7, 0.0_2_5_6] )
# fmt: on
self.assertTrue(torch_all_close(snake_case , snake_case , rtol=1E-2 ) )
def lowerCAmelCase_ ( self: Dict ) -> Optional[Any]:
# not required for this model
pass
| 66 | 1 |
"""simple docstring"""
import argparse
import pytorch_lightning as pl
import torch
from torch import nn
from transformers import LongformerForQuestionAnswering, LongformerModel
class lowerCamelCase ( pl.LightningModule ):
'''simple docstring'''
def __init__( self: str , snake_case: List[str] ) -> Tuple:
super().__init__()
snake_case_ :Any = model
snake_case_ :List[Any] = 2
snake_case_ :str = nn.Linear(self.model.config.hidden_size , self.num_labels )
def lowerCAmelCase_ ( self: str ) -> Dict:
pass
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :Optional[Any] = LongformerModel.from_pretrained(_lowercase )
snake_case_ :Dict = LightningModel(_lowercase )
snake_case_ :int = torch.load(_lowercase, map_location=torch.device("""cpu""" ) )
lightning_model.load_state_dict(ckpt["""state_dict"""] )
# init longformer question answering model
snake_case_ :List[str] = LongformerForQuestionAnswering.from_pretrained(_lowercase )
# transfer weights
longformer_for_qa.longformer.load_state_dict(lightning_model.model.state_dict() )
longformer_for_qa.qa_outputs.load_state_dict(lightning_model.qa_outputs.state_dict() )
longformer_for_qa.eval()
# save model
longformer_for_qa.save_pretrained(_lowercase )
print(f"""Conversion successful. Model saved under {pytorch_dump_folder_path}""" )
if __name__ == "__main__":
__a = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--longformer_model",
default=None,
type=str,
required=True,
help="model identifier of longformer. Should be either `longformer-base-4096` or `longformer-large-4096`.",
)
parser.add_argument(
"--longformer_question_answering_ckpt_path",
default=None,
type=str,
required=True,
help="Path the official PyTorch Lightning Checkpoint.",
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, required=True, help="Path to the output PyTorch model."
)
__a = parser.parse_args()
convert_longformer_qa_checkpoint_to_pytorch(
args.longformer_model, args.longformer_question_answering_ckpt_path, args.pytorch_dump_folder_path
)
| 66 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available, is_vision_available
__a = {
"configuration_mask2former": [
"MASK2FORMER_PRETRAINED_CONFIG_ARCHIVE_MAP",
"Mask2FormerConfig",
],
}
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["Mask2FormerImageProcessor"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"MASK2FORMER_PRETRAINED_MODEL_ARCHIVE_LIST",
"Mask2FormerForUniversalSegmentation",
"Mask2FormerModel",
"Mask2FormerPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_maskaformer import MASK2FORMER_PRETRAINED_CONFIG_ARCHIVE_MAP, MaskaFormerConfig
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .image_processing_maskaformer import MaskaFormerImageProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_maskaformer import (
MASK2FORMER_PRETRAINED_MODEL_ARCHIVE_LIST,
MaskaFormerForUniversalSegmentation,
MaskaFormerModel,
MaskaFormerPreTrainedModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure)
| 66 | 1 |
"""simple docstring"""
import tempfile
import torch
from diffusers import (
DEISMultistepScheduler,
DPMSolverMultistepScheduler,
DPMSolverSinglestepScheduler,
UniPCMultistepScheduler,
)
from .test_schedulers import SchedulerCommonTest
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
_A : Tuple = (UniPCMultistepScheduler,)
_A : Any = (("""num_inference_steps""", 2_5),)
def lowerCAmelCase_ ( self: Any , **snake_case: Optional[Any] ) -> Optional[int]:
snake_case_ :Tuple = {
"""num_train_timesteps""": 1_000,
"""beta_start""": 0.0_0_0_1,
"""beta_end""": 0.0_2,
"""beta_schedule""": """linear""",
"""solver_order""": 2,
"""solver_type""": """bh2""",
}
config.update(**snake_case )
return config
def lowerCAmelCase_ ( self: List[Any] , snake_case: Dict=0 , **snake_case: List[str] ) -> Union[str, Any]:
snake_case_ :Any = dict(self.forward_default_kwargs )
snake_case_ :str = kwargs.pop("""num_inference_steps""" , snake_case )
snake_case_ :Optional[int] = self.dummy_sample
snake_case_ :Tuple = 0.1 * sample
snake_case_ :str = [residual + 0.2, residual + 0.1_5, residual + 0.1_0]
for scheduler_class in self.scheduler_classes:
snake_case_ :Union[str, Any] = self.get_scheduler_config(**snake_case )
snake_case_ :Tuple = scheduler_class(**snake_case )
scheduler.set_timesteps(snake_case )
# copy over dummy past residuals
snake_case_ :Optional[int] = dummy_past_residuals[: scheduler.config.solver_order]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(snake_case )
snake_case_ :int = scheduler_class.from_pretrained(snake_case )
new_scheduler.set_timesteps(snake_case )
# copy over dummy past residuals
snake_case_ :Union[str, Any] = dummy_past_residuals[: new_scheduler.config.solver_order]
snake_case_, snake_case_ :Union[str, Any] = sample, sample
for t in range(snake_case , time_step + scheduler.config.solver_order + 1 ):
snake_case_ :Any = scheduler.step(snake_case , snake_case , snake_case , **snake_case ).prev_sample
snake_case_ :Dict = new_scheduler.step(snake_case , snake_case , snake_case , **snake_case ).prev_sample
assert torch.sum(torch.abs(output - new_output ) ) < 1E-5, "Scheduler outputs are not identical"
def lowerCAmelCase_ ( self: Tuple , snake_case: List[Any]=0 , **snake_case: Union[str, Any] ) -> Union[str, Any]:
snake_case_ :Dict = dict(self.forward_default_kwargs )
snake_case_ :List[Any] = kwargs.pop("""num_inference_steps""" , snake_case )
snake_case_ :Optional[int] = self.dummy_sample
snake_case_ :Any = 0.1 * sample
snake_case_ :Union[str, Any] = [residual + 0.2, residual + 0.1_5, residual + 0.1_0]
for scheduler_class in self.scheduler_classes:
snake_case_ :Tuple = self.get_scheduler_config()
snake_case_ :Dict = scheduler_class(**snake_case )
scheduler.set_timesteps(snake_case )
# copy over dummy past residuals (must be after setting timesteps)
snake_case_ :Dict = dummy_past_residuals[: scheduler.config.solver_order]
with tempfile.TemporaryDirectory() as tmpdirname:
scheduler.save_config(snake_case )
snake_case_ :Tuple = scheduler_class.from_pretrained(snake_case )
# copy over dummy past residuals
new_scheduler.set_timesteps(snake_case )
# copy over dummy past residual (must be after setting timesteps)
snake_case_ :List[Any] = dummy_past_residuals[: new_scheduler.config.solver_order]
snake_case_ :List[Any] = scheduler.step(snake_case , snake_case , snake_case , **snake_case ).prev_sample
snake_case_ :List[Any] = new_scheduler.step(snake_case , snake_case , snake_case , **snake_case ).prev_sample
assert torch.sum(torch.abs(output - new_output ) ) < 1E-5, "Scheduler outputs are not identical"
def lowerCAmelCase_ ( self: Any , snake_case: List[str]=None , **snake_case: int ) -> Any:
if scheduler is None:
snake_case_ :Optional[int] = self.scheduler_classes[0]
snake_case_ :Optional[int] = self.get_scheduler_config(**snake_case )
snake_case_ :int = scheduler_class(**snake_case )
snake_case_ :Optional[int] = self.scheduler_classes[0]
snake_case_ :List[str] = self.get_scheduler_config(**snake_case )
snake_case_ :Optional[int] = scheduler_class(**snake_case )
snake_case_ :Dict = 10
snake_case_ :List[str] = self.dummy_model()
snake_case_ :str = self.dummy_sample_deter
scheduler.set_timesteps(snake_case )
for i, t in enumerate(scheduler.timesteps ):
snake_case_ :Dict = model(snake_case , snake_case )
snake_case_ :int = scheduler.step(snake_case , snake_case , snake_case ).prev_sample
return sample
def lowerCAmelCase_ ( self: Optional[Any] ) -> Optional[int]:
snake_case_ :Dict = dict(self.forward_default_kwargs )
snake_case_ :List[Any] = kwargs.pop("""num_inference_steps""" , snake_case )
for scheduler_class in self.scheduler_classes:
snake_case_ :Union[str, Any] = self.get_scheduler_config()
snake_case_ :List[str] = scheduler_class(**snake_case )
snake_case_ :int = self.dummy_sample
snake_case_ :Tuple = 0.1 * sample
if num_inference_steps is not None and hasattr(snake_case , """set_timesteps""" ):
scheduler.set_timesteps(snake_case )
elif num_inference_steps is not None and not hasattr(snake_case , """set_timesteps""" ):
snake_case_ :Optional[int] = num_inference_steps
# copy over dummy past residuals (must be done after set_timesteps)
snake_case_ :Optional[int] = [residual + 0.2, residual + 0.1_5, residual + 0.1_0]
snake_case_ :int = dummy_past_residuals[: scheduler.config.solver_order]
snake_case_ :int = scheduler.timesteps[5]
snake_case_ :List[str] = scheduler.timesteps[6]
snake_case_ :Dict = scheduler.step(snake_case , snake_case , snake_case , **snake_case ).prev_sample
snake_case_ :Any = scheduler.step(snake_case , snake_case , snake_case , **snake_case ).prev_sample
self.assertEqual(output_a.shape , sample.shape )
self.assertEqual(output_a.shape , output_a.shape )
def lowerCAmelCase_ ( self: int ) -> List[Any]:
# make sure that iterating over schedulers with same config names gives same results
# for defaults
snake_case_ :str = UniPCMultistepScheduler(**self.get_scheduler_config() )
snake_case_ :List[Any] = self.full_loop(scheduler=snake_case )
snake_case_ :Optional[int] = torch.mean(torch.abs(snake_case ) )
assert abs(result_mean.item() - 0.2_4_6_4 ) < 1E-3
snake_case_ :Optional[int] = DPMSolverSinglestepScheduler.from_config(scheduler.config )
snake_case_ :Union[str, Any] = DEISMultistepScheduler.from_config(scheduler.config )
snake_case_ :int = DPMSolverMultistepScheduler.from_config(scheduler.config )
snake_case_ :Optional[Any] = UniPCMultistepScheduler.from_config(scheduler.config )
snake_case_ :str = self.full_loop(scheduler=snake_case )
snake_case_ :Dict = torch.mean(torch.abs(snake_case ) )
assert abs(result_mean.item() - 0.2_4_6_4 ) < 1E-3
def lowerCAmelCase_ ( self: Any ) -> List[str]:
for timesteps in [25, 50, 100, 999, 1_000]:
self.check_over_configs(num_train_timesteps=snake_case )
def lowerCAmelCase_ ( self: Tuple ) -> Tuple:
self.check_over_configs(thresholding=snake_case )
for order in [1, 2, 3]:
for solver_type in ["bh1", "bh2"]:
for threshold in [0.5, 1.0, 2.0]:
for prediction_type in ["epsilon", "sample"]:
self.check_over_configs(
thresholding=snake_case , prediction_type=snake_case , sample_max_value=snake_case , solver_order=snake_case , solver_type=snake_case , )
def lowerCAmelCase_ ( self: Dict ) -> List[str]:
for prediction_type in ["epsilon", "v_prediction"]:
self.check_over_configs(prediction_type=snake_case )
def lowerCAmelCase_ ( self: str ) -> Optional[int]:
for solver_type in ["bh1", "bh2"]:
for order in [1, 2, 3]:
for prediction_type in ["epsilon", "sample"]:
self.check_over_configs(
solver_order=snake_case , solver_type=snake_case , prediction_type=snake_case , )
snake_case_ :str = self.full_loop(
solver_order=snake_case , solver_type=snake_case , prediction_type=snake_case , )
assert not torch.isnan(snake_case ).any(), "Samples have nan numbers"
def lowerCAmelCase_ ( self: List[Any] ) -> Union[str, Any]:
self.check_over_configs(lower_order_final=snake_case )
self.check_over_configs(lower_order_final=snake_case )
def lowerCAmelCase_ ( self: List[str] ) -> int:
for num_inference_steps in [1, 2, 3, 5, 10, 50, 100, 999, 1_000]:
self.check_over_forward(num_inference_steps=snake_case , time_step=0 )
def lowerCAmelCase_ ( self: str ) -> Optional[int]:
snake_case_ :Tuple = self.full_loop()
snake_case_ :List[Any] = torch.mean(torch.abs(snake_case ) )
assert abs(result_mean.item() - 0.2_4_6_4 ) < 1E-3
def lowerCAmelCase_ ( self: Tuple ) -> Optional[int]:
snake_case_ :Dict = self.full_loop(prediction_type="""v_prediction""" )
snake_case_ :Any = torch.mean(torch.abs(snake_case ) )
assert abs(result_mean.item() - 0.1_0_1_4 ) < 1E-3
def lowerCAmelCase_ ( self: Any ) -> List[str]:
snake_case_ :Dict = self.scheduler_classes[0]
snake_case_ :Tuple = self.get_scheduler_config(thresholding=snake_case , dynamic_thresholding_ratio=0 )
snake_case_ :List[Any] = scheduler_class(**snake_case )
snake_case_ :Any = 10
snake_case_ :int = self.dummy_model()
snake_case_ :Union[str, Any] = self.dummy_sample_deter.half()
scheduler.set_timesteps(snake_case )
for i, t in enumerate(scheduler.timesteps ):
snake_case_ :int = model(snake_case , snake_case )
snake_case_ :List[Any] = scheduler.step(snake_case , snake_case , snake_case ).prev_sample
assert sample.dtype == torch.floataa
def lowerCAmelCase_ ( self: Tuple , **snake_case: List[Any] ) -> Dict:
for scheduler_class in self.scheduler_classes:
snake_case_ :int = self.get_scheduler_config(**snake_case )
snake_case_ :Tuple = scheduler_class(**snake_case )
scheduler.set_timesteps(scheduler.config.num_train_timesteps )
assert len(scheduler.timesteps.unique() ) == scheduler.num_inference_steps
| 66 |
"""simple docstring"""
import gc
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import (
AutoencoderKL,
DDIMScheduler,
StableDiffusionSAGPipeline,
UNetaDConditionModel,
)
from diffusers.utils import slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu
from ..pipeline_params import TEXT_TO_IMAGE_BATCH_PARAMS, TEXT_TO_IMAGE_IMAGE_PARAMS, TEXT_TO_IMAGE_PARAMS
from ..test_pipelines_common import PipelineLatentTesterMixin, PipelineTesterMixin
enable_full_determinism()
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : str = StableDiffusionSAGPipeline
_A : Optional[Any] = TEXT_TO_IMAGE_PARAMS
_A : Any = TEXT_TO_IMAGE_BATCH_PARAMS
_A : Tuple = TEXT_TO_IMAGE_IMAGE_PARAMS
_A : Tuple = TEXT_TO_IMAGE_IMAGE_PARAMS
_A : List[str] = False
def lowerCAmelCase_ ( self: Optional[Any] ) -> str:
torch.manual_seed(0 )
snake_case_ :Any = UNetaDConditionModel(
block_out_channels=(32, 64) , layers_per_block=2 , sample_size=32 , in_channels=4 , out_channels=4 , down_block_types=("""DownBlock2D""", """CrossAttnDownBlock2D""") , up_block_types=("""CrossAttnUpBlock2D""", """UpBlock2D""") , cross_attention_dim=32 , )
snake_case_ :Any = DDIMScheduler(
beta_start=0.0_0_0_8_5 , beta_end=0.0_1_2 , beta_schedule="""scaled_linear""" , clip_sample=snake_case , set_alpha_to_one=snake_case , )
torch.manual_seed(0 )
snake_case_ :Optional[int] = AutoencoderKL(
block_out_channels=[32, 64] , in_channels=3 , out_channels=3 , down_block_types=["""DownEncoderBlock2D""", """DownEncoderBlock2D"""] , up_block_types=["""UpDecoderBlock2D""", """UpDecoderBlock2D"""] , latent_channels=4 , )
torch.manual_seed(0 )
snake_case_ :Union[str, Any] = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1E-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1_000 , )
snake_case_ :Tuple = CLIPTextModel(snake_case )
snake_case_ :str = CLIPTokenizer.from_pretrained("""hf-internal-testing/tiny-random-clip""" )
snake_case_ :Dict = {
"""unet""": unet,
"""scheduler""": scheduler,
"""vae""": vae,
"""text_encoder""": text_encoder,
"""tokenizer""": tokenizer,
"""safety_checker""": None,
"""feature_extractor""": None,
}
return components
def lowerCAmelCase_ ( self: List[str] , snake_case: Tuple , snake_case: List[str]=0 ) -> str:
if str(snake_case ).startswith("""mps""" ):
snake_case_ :Tuple = torch.manual_seed(snake_case )
else:
snake_case_ :Optional[int] = torch.Generator(device=snake_case ).manual_seed(snake_case )
snake_case_ :Any = {
"""prompt""": """.""",
"""generator""": generator,
"""num_inference_steps""": 2,
"""guidance_scale""": 1.0,
"""sag_scale""": 1.0,
"""output_type""": """numpy""",
}
return inputs
def lowerCAmelCase_ ( self: Optional[int] ) -> str:
super().test_inference_batch_single_identical(expected_max_diff=3E-3 )
@slow
@require_torch_gpu
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: int ) -> str:
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def lowerCAmelCase_ ( self: int ) -> List[str]:
snake_case_ :Any = StableDiffusionSAGPipeline.from_pretrained("""CompVis/stable-diffusion-v1-4""" )
snake_case_ :int = sag_pipe.to(snake_case )
sag_pipe.set_progress_bar_config(disable=snake_case )
snake_case_ :Union[str, Any] = """."""
snake_case_ :str = torch.manual_seed(0 )
snake_case_ :str = sag_pipe(
[prompt] , generator=snake_case , guidance_scale=7.5 , sag_scale=1.0 , num_inference_steps=20 , output_type="""np""" )
snake_case_ :List[Any] = output.images
snake_case_ :Tuple = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
snake_case_ :List[Any] = np.array([0.1_5_6_8, 0.1_7_3_8, 0.1_6_9_5, 0.1_6_9_3, 0.1_5_0_7, 0.1_7_0_5, 0.1_5_4_7, 0.1_7_5_1, 0.1_9_4_9] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 5E-2
def lowerCAmelCase_ ( self: Dict ) -> str:
snake_case_ :Tuple = StableDiffusionSAGPipeline.from_pretrained("""stabilityai/stable-diffusion-2-1-base""" )
snake_case_ :Optional[int] = sag_pipe.to(snake_case )
sag_pipe.set_progress_bar_config(disable=snake_case )
snake_case_ :Tuple = """."""
snake_case_ :Union[str, Any] = torch.manual_seed(0 )
snake_case_ :Tuple = sag_pipe(
[prompt] , generator=snake_case , guidance_scale=7.5 , sag_scale=1.0 , num_inference_steps=20 , output_type="""np""" )
snake_case_ :Optional[int] = output.images
snake_case_ :Optional[Any] = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
snake_case_ :Tuple = np.array([0.3_4_5_9, 0.2_8_7_6, 0.2_5_3_7, 0.3_0_0_2, 0.2_6_7_1, 0.2_1_6_0, 0.3_0_2_6, 0.2_2_6_2, 0.2_3_7_1] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 5E-2
def lowerCAmelCase_ ( self: List[str] ) -> List[str]:
snake_case_ :Optional[int] = StableDiffusionSAGPipeline.from_pretrained("""stabilityai/stable-diffusion-2-1-base""" )
snake_case_ :int = sag_pipe.to(snake_case )
sag_pipe.set_progress_bar_config(disable=snake_case )
snake_case_ :Tuple = """."""
snake_case_ :Optional[int] = torch.manual_seed(0 )
snake_case_ :List[str] = sag_pipe(
[prompt] , width=768 , height=512 , generator=snake_case , guidance_scale=7.5 , sag_scale=1.0 , num_inference_steps=20 , output_type="""np""" , )
snake_case_ :Optional[Any] = output.images
assert image.shape == (1, 512, 768, 3)
| 66 | 1 |
"""simple docstring"""
from ..utils import (
OptionalDependencyNotAvailable,
is_flax_available,
is_scipy_available,
is_torch_available,
is_torchsde_available,
)
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from ..utils.dummy_pt_objects import * # noqa F403
else:
from .scheduling_consistency_models import CMStochasticIterativeScheduler
from .scheduling_ddim import DDIMScheduler
from .scheduling_ddim_inverse import DDIMInverseScheduler
from .scheduling_ddim_parallel import DDIMParallelScheduler
from .scheduling_ddpm import DDPMScheduler
from .scheduling_ddpm_parallel import DDPMParallelScheduler
from .scheduling_deis_multistep import DEISMultistepScheduler
from .scheduling_dpmsolver_multistep import DPMSolverMultistepScheduler
from .scheduling_dpmsolver_multistep_inverse import DPMSolverMultistepInverseScheduler
from .scheduling_dpmsolver_singlestep import DPMSolverSinglestepScheduler
from .scheduling_euler_ancestral_discrete import EulerAncestralDiscreteScheduler
from .scheduling_euler_discrete import EulerDiscreteScheduler
from .scheduling_heun_discrete import HeunDiscreteScheduler
from .scheduling_ipndm import IPNDMScheduler
from .scheduling_k_dpm_2_ancestral_discrete import KDPMaAncestralDiscreteScheduler
from .scheduling_k_dpm_2_discrete import KDPMaDiscreteScheduler
from .scheduling_karras_ve import KarrasVeScheduler
from .scheduling_pndm import PNDMScheduler
from .scheduling_repaint import RePaintScheduler
from .scheduling_sde_ve import ScoreSdeVeScheduler
from .scheduling_sde_vp import ScoreSdeVpScheduler
from .scheduling_unclip import UnCLIPScheduler
from .scheduling_unipc_multistep import UniPCMultistepScheduler
from .scheduling_utils import KarrasDiffusionSchedulers, SchedulerMixin
from .scheduling_vq_diffusion import VQDiffusionScheduler
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from ..utils.dummy_flax_objects import * # noqa F403
else:
from .scheduling_ddim_flax import FlaxDDIMScheduler
from .scheduling_ddpm_flax import FlaxDDPMScheduler
from .scheduling_dpmsolver_multistep_flax import FlaxDPMSolverMultistepScheduler
from .scheduling_karras_ve_flax import FlaxKarrasVeScheduler
from .scheduling_lms_discrete_flax import FlaxLMSDiscreteScheduler
from .scheduling_pndm_flax import FlaxPNDMScheduler
from .scheduling_sde_ve_flax import FlaxScoreSdeVeScheduler
from .scheduling_utils_flax import (
FlaxKarrasDiffusionSchedulers,
FlaxSchedulerMixin,
FlaxSchedulerOutput,
broadcast_to_shape_from_left,
)
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 .scheduling_lms_discrete 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 .scheduling_dpmsolver_sde import DPMSolverSDEScheduler
| 66 |
"""simple docstring"""
from __future__ import annotations
from collections import Counter
from random import random
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Tuple ) -> Optional[Any]:
snake_case_ :Optional[int] = {}
def lowerCAmelCase_ ( self: Dict , snake_case: str ) -> None:
snake_case_ :str = {}
def lowerCAmelCase_ ( self: Optional[int] , snake_case: str , snake_case: str , snake_case: float ) -> None:
if nodea not in self.connections:
self.add_node(snake_case )
if nodea not in self.connections:
self.add_node(snake_case )
snake_case_ :Dict = probability
def lowerCAmelCase_ ( self: List[Any] ) -> list[str]:
return list(self.connections )
def lowerCAmelCase_ ( self: Any , snake_case: str ) -> str:
snake_case_ :Optional[Any] = 0
snake_case_ :List[str] = random()
for dest in self.connections[node]:
current_probability += self.connections[node][dest]
if current_probability > random_value:
return dest
return ""
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :List[str] = MarkovChainGraphUndirectedUnweighted()
for nodea, nodea, probability in transitions:
graph.add_transition_probability(_lowercase, _lowercase, _lowercase )
snake_case_ :int = Counter(graph.get_nodes() )
snake_case_ :Optional[Any] = start
for _ in range(_lowercase ):
snake_case_ :Tuple = graph.transition(_lowercase )
visited[node] += 1
return visited
if __name__ == "__main__":
import doctest
doctest.testmod()
| 66 | 1 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_sentencepiece_available,
is_torch_available,
)
__a = {
"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:
__a = ["SpeechT5Tokenizer"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"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
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 |
"""simple docstring"""
import argparse
import collections
import os
import re
import tempfile
import pandas as pd
from datasets import Dataset
from huggingface_hub import hf_hub_download, upload_folder
from transformers.utils import direct_transformers_import
# All paths are set with the intent you should run this script from the root of the repo with the command
# python utils/update_metadata.py
__a = "src/transformers"
# This is to make sure the transformers module imported is the one in the repo.
__a = direct_transformers_import(TRANSFORMERS_PATH)
# Regexes that match TF/Flax/PT model names.
__a = re.compile(r"TF(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
__a = re.compile(r"Flax(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
# Will match any TF or Flax model too so need to be in an else branch afterthe two previous regexes.
__a = re.compile(r"(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
# Fill this with tuples (pipeline_tag, model_mapping, auto_model)
__a = [
("pretraining", "MODEL_FOR_PRETRAINING_MAPPING_NAMES", "AutoModelForPreTraining"),
("feature-extraction", "MODEL_MAPPING_NAMES", "AutoModel"),
("audio-classification", "MODEL_FOR_AUDIO_CLASSIFICATION_MAPPING_NAMES", "AutoModelForAudioClassification"),
("text-generation", "MODEL_FOR_CAUSAL_LM_MAPPING_NAMES", "AutoModelForCausalLM"),
("automatic-speech-recognition", "MODEL_FOR_CTC_MAPPING_NAMES", "AutoModelForCTC"),
("image-classification", "MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES", "AutoModelForImageClassification"),
("image-segmentation", "MODEL_FOR_IMAGE_SEGMENTATION_MAPPING_NAMES", "AutoModelForImageSegmentation"),
("fill-mask", "MODEL_FOR_MASKED_LM_MAPPING_NAMES", "AutoModelForMaskedLM"),
("object-detection", "MODEL_FOR_OBJECT_DETECTION_MAPPING_NAMES", "AutoModelForObjectDetection"),
(
"zero-shot-object-detection",
"MODEL_FOR_ZERO_SHOT_OBJECT_DETECTION_MAPPING_NAMES",
"AutoModelForZeroShotObjectDetection",
),
("question-answering", "MODEL_FOR_QUESTION_ANSWERING_MAPPING_NAMES", "AutoModelForQuestionAnswering"),
("text2text-generation", "MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING_NAMES", "AutoModelForSeq2SeqLM"),
("text-classification", "MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES", "AutoModelForSequenceClassification"),
("automatic-speech-recognition", "MODEL_FOR_SPEECH_SEQ_2_SEQ_MAPPING_NAMES", "AutoModelForSpeechSeq2Seq"),
(
"table-question-answering",
"MODEL_FOR_TABLE_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForTableQuestionAnswering",
),
("token-classification", "MODEL_FOR_TOKEN_CLASSIFICATION_MAPPING_NAMES", "AutoModelForTokenClassification"),
("multiple-choice", "MODEL_FOR_MULTIPLE_CHOICE_MAPPING_NAMES", "AutoModelForMultipleChoice"),
(
"next-sentence-prediction",
"MODEL_FOR_NEXT_SENTENCE_PREDICTION_MAPPING_NAMES",
"AutoModelForNextSentencePrediction",
),
(
"audio-frame-classification",
"MODEL_FOR_AUDIO_FRAME_CLASSIFICATION_MAPPING_NAMES",
"AutoModelForAudioFrameClassification",
),
("audio-xvector", "MODEL_FOR_AUDIO_XVECTOR_MAPPING_NAMES", "AutoModelForAudioXVector"),
(
"document-question-answering",
"MODEL_FOR_DOCUMENT_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForDocumentQuestionAnswering",
),
(
"visual-question-answering",
"MODEL_FOR_VISUAL_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForVisualQuestionAnswering",
),
("image-to-text", "MODEL_FOR_FOR_VISION_2_SEQ_MAPPING_NAMES", "AutoModelForVision2Seq"),
(
"zero-shot-image-classification",
"MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING_NAMES",
"AutoModelForZeroShotImageClassification",
),
("depth-estimation", "MODEL_FOR_DEPTH_ESTIMATION_MAPPING_NAMES", "AutoModelForDepthEstimation"),
("video-classification", "MODEL_FOR_VIDEO_CLASSIFICATION_MAPPING_NAMES", "AutoModelForVideoClassification"),
("mask-generation", "MODEL_FOR_MASK_GENERATION_MAPPING_NAMES", "AutoModelForMaskGeneration"),
]
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Any = re.finditer(""".+?(?:(?<=[a-z])(?=[A-Z])|(?<=[A-Z])(?=[A-Z][a-z])|$)""", _lowercase )
return [m.group(0 ) for m in matches]
def A_ ( ):
'''simple docstring'''
snake_case_ :int = transformers_module.models.auto.configuration_auto.CONFIG_MAPPING_NAMES
snake_case_ :Dict = {
config.replace("""Config""", """""" ): model_type for model_type, config in config_maping_names.items()
}
# Dictionaries flagging if each model prefix has a backend in PT/TF/Flax.
snake_case_ :Optional[Any] = collections.defaultdict(_lowercase )
snake_case_ :int = collections.defaultdict(_lowercase )
snake_case_ :List[str] = collections.defaultdict(_lowercase )
# Let's lookup through all transformers object (once) and find if models are supported by a given backend.
for attr_name in dir(_lowercase ):
snake_case_ :int = None
if _re_tf_models.match(_lowercase ) is not None:
snake_case_ :int = tf_models
snake_case_ :List[str] = _re_tf_models.match(_lowercase ).groups()[0]
elif _re_flax_models.match(_lowercase ) is not None:
snake_case_ :List[Any] = flax_models
snake_case_ :Any = _re_flax_models.match(_lowercase ).groups()[0]
elif _re_pt_models.match(_lowercase ) is not None:
snake_case_ :Optional[Any] = pt_models
snake_case_ :int = _re_pt_models.match(_lowercase ).groups()[0]
if lookup_dict is not None:
while len(_lowercase ) > 0:
if attr_name in model_prefix_to_model_type:
snake_case_ :Optional[int] = True
break
# Try again after removing the last word in the name
snake_case_ :Optional[Any] = """""".join(camel_case_split(_lowercase )[:-1] )
snake_case_ :Optional[int] = set(list(pt_models.keys() ) + list(tf_models.keys() ) + list(flax_models.keys() ) )
snake_case_ :Optional[Any] = list(_lowercase )
all_models.sort()
snake_case_ :Optional[int] = {"""model_type""": all_models}
snake_case_ :Optional[int] = [pt_models[t] for t in all_models]
snake_case_ :Any = [tf_models[t] for t in all_models]
snake_case_ :Dict = [flax_models[t] for t in all_models]
# Now let's use the auto-mapping names to make sure
snake_case_ :Dict = {}
for t in all_models:
if t in transformers_module.models.auto.processing_auto.PROCESSOR_MAPPING_NAMES:
snake_case_ :Optional[Any] = """AutoProcessor"""
elif t in transformers_module.models.auto.tokenization_auto.TOKENIZER_MAPPING_NAMES:
snake_case_ :Tuple = """AutoTokenizer"""
elif t in transformers_module.models.auto.feature_extraction_auto.FEATURE_EXTRACTOR_MAPPING_NAMES:
snake_case_ :Tuple = """AutoFeatureExtractor"""
else:
# Default to AutoTokenizer if a model has nothing, for backward compatibility.
snake_case_ :str = """AutoTokenizer"""
snake_case_ :int = [processors[t] for t in all_models]
return pd.DataFrame(_lowercase )
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :List[Any] = [
transformers_module.models.auto.modeling_auto,
transformers_module.models.auto.modeling_tf_auto,
transformers_module.models.auto.modeling_flax_auto,
]
for pipeline_tag, model_mapping, auto_class in PIPELINE_TAGS_AND_AUTO_MODELS:
snake_case_ :Optional[int] = [model_mapping, f"""TF_{model_mapping}""", f"""FLAX_{model_mapping}"""]
snake_case_ :List[str] = [auto_class, f"""TF_{auto_class}""", f"""Flax_{auto_class}"""]
# Loop through all three frameworks
for module, cls, mapping in zip(_lowercase, _lowercase, _lowercase ):
# The type of pipeline may not exist in this framework
if not hasattr(_lowercase, _lowercase ):
continue
# First extract all model_names
snake_case_ :Tuple = []
for name in getattr(_lowercase, _lowercase ).values():
if isinstance(_lowercase, _lowercase ):
model_names.append(_lowercase )
else:
model_names.extend(list(_lowercase ) )
# Add pipeline tag and auto model class for those models
table.update({model_name: (pipeline_tag, cls) for model_name in model_names} )
return table
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :List[Any] = get_frameworks_table()
snake_case_ :str = Dataset.from_pandas(_lowercase )
snake_case_ :List[Any] = hf_hub_download(
"""huggingface/transformers-metadata""", """pipeline_tags.json""", repo_type="""dataset""", token=_lowercase )
snake_case_ :List[str] = Dataset.from_json(_lowercase )
snake_case_ :int = {
tags_dataset[i]["""model_class"""]: (tags_dataset[i]["""pipeline_tag"""], tags_dataset[i]["""auto_class"""])
for i in range(len(_lowercase ) )
}
snake_case_ :Optional[int] = update_pipeline_and_auto_class_table(_lowercase )
# Sort the model classes to avoid some nondeterministic updates to create false update commits.
snake_case_ :Tuple = sorted(table.keys() )
snake_case_ :Tuple = pd.DataFrame(
{
"""model_class""": model_classes,
"""pipeline_tag""": [table[m][0] for m in model_classes],
"""auto_class""": [table[m][1] for m in model_classes],
} )
snake_case_ :Union[str, Any] = Dataset.from_pandas(_lowercase )
with tempfile.TemporaryDirectory() as tmp_dir:
frameworks_dataset.to_json(os.path.join(_lowercase, """frameworks.json""" ) )
tags_dataset.to_json(os.path.join(_lowercase, """pipeline_tags.json""" ) )
if commit_sha is not None:
snake_case_ :Union[str, Any] = (
f"""Update with commit {commit_sha}\n\nSee: """
f"""https://github.com/huggingface/transformers/commit/{commit_sha}"""
)
else:
snake_case_ :List[Any] = """Update"""
upload_folder(
repo_id="""huggingface/transformers-metadata""", folder_path=_lowercase, repo_type="""dataset""", token=_lowercase, commit_message=_lowercase, )
def A_ ( ):
'''simple docstring'''
snake_case_ :List[Any] = {tag: cls for tag, _, cls in PIPELINE_TAGS_AND_AUTO_MODELS}
snake_case_ :Dict = transformers_module.pipelines.SUPPORTED_TASKS
snake_case_ :List[str] = []
for key in pipeline_tasks:
if key not in in_table:
snake_case_ :int = pipeline_tasks[key]["""pt"""]
if isinstance(_lowercase, (list, tuple) ):
snake_case_ :Any = model[0]
snake_case_ :str = model.__name__
if model not in in_table.values():
missing.append(_lowercase )
if len(_lowercase ) > 0:
snake_case_ :Optional[int] = """, """.join(_lowercase )
raise ValueError(
"""The following pipeline tags are not present in the `PIPELINE_TAGS_AND_AUTO_MODELS` constant inside """
f"""`utils/update_metadata.py`: {msg}. Please add them!""" )
if __name__ == "__main__":
__a = argparse.ArgumentParser()
parser.add_argument("--token", type=str, help="The token to use to push to the transformers-metadata dataset.")
parser.add_argument("--commit_sha", type=str, help="The sha of the commit going with this update.")
parser.add_argument("--check-only", action="store_true", help="Activate to just check all pipelines are present.")
__a = parser.parse_args()
if args.check_only:
check_pipeline_tags()
else:
update_metadata(args.token, args.commit_sha)
| 66 | 1 |
"""simple docstring"""
from __future__ import annotations
__a = tuple[int, int, int]
__a = tuple[str, str, str]
# used alphabet --------------------------
# from string.ascii_uppercase
__a = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
# -------------------------- default selection --------------------------
# rotors --------------------------
__a = "EGZWVONAHDCLFQMSIPJBYUKXTR"
__a = "FOBHMDKEXQNRAULPGSJVTYICZW"
__a = "ZJXESIUQLHAVRMDOYGTNFWPBKC"
# reflector --------------------------
__a = {
"A": "N",
"N": "A",
"B": "O",
"O": "B",
"C": "P",
"P": "C",
"D": "Q",
"Q": "D",
"E": "R",
"R": "E",
"F": "S",
"S": "F",
"G": "T",
"T": "G",
"H": "U",
"U": "H",
"I": "V",
"V": "I",
"J": "W",
"W": "J",
"K": "X",
"X": "K",
"L": "Y",
"Y": "L",
"M": "Z",
"Z": "M",
}
# -------------------------- extra rotors --------------------------
__a = "RMDJXFUWGISLHVTCQNKYPBEZOA"
__a = "SGLCPQWZHKXAREONTFBVIYJUDM"
__a = "HVSICLTYKQUBXDWAJZOMFGPREN"
__a = "RZWQHFMVDBKICJLNTUXAGYPSOE"
__a = "LFKIJODBEGAMQPXVUHYSTCZRWN"
__a = "KOAEGVDHXPQZMLFTYWJNBRCIUS"
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
if (unique_rotsel := len(set(_lowercase ) )) < 3:
snake_case_ :Any = f"""Please use 3 unique rotors (not {unique_rotsel})"""
raise Exception(_lowercase )
# Checks if rotor positions are valid
snake_case_, snake_case_, snake_case_ :int = rotpos
if not 0 < rotorposa <= len(_lowercase ):
snake_case_ :List[Any] = f"""First rotor position is not within range of 1..26 ({rotorposa}"""
raise ValueError(_lowercase )
if not 0 < rotorposa <= len(_lowercase ):
snake_case_ :Tuple = f"""Second rotor position is not within range of 1..26 ({rotorposa})"""
raise ValueError(_lowercase )
if not 0 < rotorposa <= len(_lowercase ):
snake_case_ :str = f"""Third rotor position is not within range of 1..26 ({rotorposa})"""
raise ValueError(_lowercase )
# Validates string and returns dict
snake_case_ :Optional[Any] = _plugboard(_lowercase )
return rotpos, rotsel, pbdict
def A_ ( _lowercase ):
'''simple docstring'''
if not isinstance(_lowercase, _lowercase ):
snake_case_ :int = f"""Plugboard setting isn't type string ({type(_lowercase )})"""
raise TypeError(_lowercase )
elif len(_lowercase ) % 2 != 0:
snake_case_ :List[Any] = f"""Odd number of symbols ({len(_lowercase )})"""
raise Exception(_lowercase )
elif pbstring == "":
return {}
pbstring.replace(""" """, """""" )
# Checks if all characters are unique
snake_case_ :List[str] = set()
for i in pbstring:
if i not in abc:
snake_case_ :Dict = f"""'{i}' not in list of symbols"""
raise Exception(_lowercase )
elif i in tmppbl:
snake_case_ :Dict = f"""Duplicate symbol ({i})"""
raise Exception(_lowercase )
else:
tmppbl.add(_lowercase )
del tmppbl
# Created the dictionary
snake_case_ :int = {}
for j in range(0, len(_lowercase ) - 1, 2 ):
snake_case_ :Dict = pbstring[j + 1]
snake_case_ :List[Any] = pbstring[j]
return pb
def A_ ( _lowercase, _lowercase, _lowercase = (rotora, rotora, rotora), _lowercase = "", ):
'''simple docstring'''
snake_case_ :Tuple = text.upper()
snake_case_, snake_case_, snake_case_ :Tuple = _validator(
_lowercase, _lowercase, plugb.upper() )
snake_case_, snake_case_, snake_case_ :int = rotor_position
snake_case_, snake_case_, snake_case_ :Tuple = rotor_selection
rotorposa -= 1
rotorposa -= 1
rotorposa -= 1
snake_case_ :int = []
# encryption/decryption process --------------------------
for symbol in text:
if symbol in abc:
# 1st plugboard --------------------------
if symbol in plugboard:
snake_case_ :Any = plugboard[symbol]
# rotor ra --------------------------
snake_case_ :Optional[int] = abc.index(_lowercase ) + rotorposa
snake_case_ :Any = rotora[index % len(_lowercase )]
# rotor rb --------------------------
snake_case_ :List[Any] = abc.index(_lowercase ) + rotorposa
snake_case_ :int = rotora[index % len(_lowercase )]
# rotor rc --------------------------
snake_case_ :int = abc.index(_lowercase ) + rotorposa
snake_case_ :List[Any] = rotora[index % len(_lowercase )]
# reflector --------------------------
# this is the reason you don't need another machine to decipher
snake_case_ :Union[str, Any] = reflector[symbol]
# 2nd rotors
snake_case_ :int = abc[rotora.index(_lowercase ) - rotorposa]
snake_case_ :Dict = abc[rotora.index(_lowercase ) - rotorposa]
snake_case_ :Union[str, Any] = abc[rotora.index(_lowercase ) - rotorposa]
# 2nd plugboard
if symbol in plugboard:
snake_case_ :int = plugboard[symbol]
# moves/resets rotor positions
rotorposa += 1
if rotorposa >= len(_lowercase ):
snake_case_ :List[Any] = 0
rotorposa += 1
if rotorposa >= len(_lowercase ):
snake_case_ :str = 0
rotorposa += 1
if rotorposa >= len(_lowercase ):
snake_case_ :Union[str, Any] = 0
# else:
# pass
# Error could be also raised
# raise ValueError(
# 'Invalid symbol('+repr(symbol)+')')
result.append(_lowercase )
return "".join(_lowercase )
if __name__ == "__main__":
__a = "This is my Python script that emulates the Enigma machine from WWII."
__a = (1, 1, 1)
__a = "pictures"
__a = (rotora, rotora, rotora)
__a = enigma(message, rotor_pos, rotor_sel, pb)
print("Encrypted message:", en)
print("Decrypted message:", enigma(en, rotor_pos, rotor_sel, pb))
| 66 |
"""simple docstring"""
import argparse
import glob
import logging
import os
from argparse import Namespace
from importlib import import_module
import numpy as np
import torch
from lightning_base import BaseTransformer, add_generic_args, generic_train
from seqeval.metrics import accuracy_score, fa_score, precision_score, recall_score
from torch.nn import CrossEntropyLoss
from torch.utils.data import DataLoader, TensorDataset
from utils_ner import TokenClassificationTask
__a = logging.getLogger(__name__)
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
_A : Union[str, Any] = """token-classification"""
def __init__( self: Any , snake_case: Tuple ) -> List[Any]:
if type(snake_case ) == dict:
snake_case_ :Optional[int] = Namespace(**snake_case )
snake_case_ :Optional[int] = import_module("""tasks""" )
try:
snake_case_ :Any = getattr(snake_case , hparams.task_type )
snake_case_ :TokenClassificationTask = token_classification_task_clazz()
except AttributeError:
raise ValueError(
f"""Task {hparams.task_type} needs to be defined as a TokenClassificationTask subclass in {module}. """
f"""Available tasks classes are: {TokenClassificationTask.__subclasses__()}""" )
snake_case_ :Any = self.token_classification_task.get_labels(hparams.labels )
snake_case_ :str = CrossEntropyLoss().ignore_index
super().__init__(snake_case , len(self.labels ) , self.mode )
def lowerCAmelCase_ ( self: Dict , **snake_case: List[Any] ) -> Any:
return self.model(**snake_case )
def lowerCAmelCase_ ( self: str , snake_case: Tuple , snake_case: List[Any] ) -> Optional[int]:
snake_case_ :List[str] = {"""input_ids""": batch[0], """attention_mask""": batch[1], """labels""": batch[3]}
if self.config.model_type != "distilbert":
snake_case_ :List[str] = (
batch[2] if self.config.model_type in ["""bert""", """xlnet"""] else None
) # XLM and RoBERTa don"t use token_type_ids
snake_case_ :Optional[Any] = self(**snake_case )
snake_case_ :List[str] = outputs[0]
# tensorboard_logs = {"loss": loss, "rate": self.lr_scheduler.get_last_lr()[-1]}
return {"loss": loss}
def lowerCAmelCase_ ( self: int ) -> Dict:
snake_case_ :List[Any] = self.hparams
for mode in ["train", "dev", "test"]:
snake_case_ :Optional[int] = self._feature_file(snake_case )
if os.path.exists(snake_case ) and not args.overwrite_cache:
logger.info("""Loading features from cached file %s""" , snake_case )
snake_case_ :Optional[int] = torch.load(snake_case )
else:
logger.info("""Creating features from dataset file at %s""" , args.data_dir )
snake_case_ :Optional[int] = self.token_classification_task.read_examples_from_file(args.data_dir , snake_case )
snake_case_ :Any = self.token_classification_task.convert_examples_to_features(
snake_case , self.labels , args.max_seq_length , self.tokenizer , cls_token_at_end=bool(self.config.model_type in ["""xlnet"""] ) , cls_token=self.tokenizer.cls_token , cls_token_segment_id=2 if self.config.model_type in ["""xlnet"""] else 0 , sep_token=self.tokenizer.sep_token , sep_token_extra=snake_case , pad_on_left=bool(self.config.model_type in ["""xlnet"""] ) , pad_token=self.tokenizer.pad_token_id , pad_token_segment_id=self.tokenizer.pad_token_type_id , pad_token_label_id=self.pad_token_label_id , )
logger.info("""Saving features into cached file %s""" , snake_case )
torch.save(snake_case , snake_case )
def lowerCAmelCase_ ( self: Optional[int] , snake_case: int , snake_case: int , snake_case: bool = False ) -> DataLoader:
snake_case_ :int = self._feature_file(snake_case )
logger.info("""Loading features from cached file %s""" , snake_case )
snake_case_ :str = torch.load(snake_case )
snake_case_ :Dict = torch.tensor([f.input_ids for f in features] , dtype=torch.long )
snake_case_ :str = torch.tensor([f.attention_mask for f in features] , dtype=torch.long )
if features[0].token_type_ids is not None:
snake_case_ :List[Any] = torch.tensor([f.token_type_ids for f in features] , dtype=torch.long )
else:
snake_case_ :List[str] = torch.tensor([0 for f in features] , dtype=torch.long )
# HACK(we will not use this anymore soon)
snake_case_ :Any = torch.tensor([f.label_ids for f in features] , dtype=torch.long )
return DataLoader(
TensorDataset(snake_case , snake_case , snake_case , snake_case ) , batch_size=snake_case )
def lowerCAmelCase_ ( self: List[str] , snake_case: Dict , snake_case: Union[str, Any] ) -> List[str]:
"""Compute validation""" ""
snake_case_ :List[str] = {"""input_ids""": batch[0], """attention_mask""": batch[1], """labels""": batch[3]}
if self.config.model_type != "distilbert":
snake_case_ :Dict = (
batch[2] if self.config.model_type in ["""bert""", """xlnet"""] else None
) # XLM and RoBERTa don"t use token_type_ids
snake_case_ :Dict = self(**snake_case )
snake_case_, snake_case_ :Dict = outputs[:2]
snake_case_ :Union[str, Any] = logits.detach().cpu().numpy()
snake_case_ :List[Any] = inputs["""labels"""].detach().cpu().numpy()
return {"val_loss": tmp_eval_loss.detach().cpu(), "pred": preds, "target": out_label_ids}
def lowerCAmelCase_ ( self: List[Any] , snake_case: int ) -> Tuple:
snake_case_ :Union[str, Any] = torch.stack([x["""val_loss"""] for x in outputs] ).mean()
snake_case_ :Tuple = np.concatenate([x["""pred"""] for x in outputs] , axis=0 )
snake_case_ :Tuple = np.argmax(snake_case , axis=2 )
snake_case_ :List[str] = np.concatenate([x["""target"""] for x in outputs] , axis=0 )
snake_case_ :Optional[Any] = dict(enumerate(self.labels ) )
snake_case_ :Dict = [[] for _ in range(out_label_ids.shape[0] )]
snake_case_ :Dict = [[] for _ in range(out_label_ids.shape[0] )]
for i in range(out_label_ids.shape[0] ):
for j in range(out_label_ids.shape[1] ):
if out_label_ids[i, j] != self.pad_token_label_id:
out_label_list[i].append(label_map[out_label_ids[i][j]] )
preds_list[i].append(label_map[preds[i][j]] )
snake_case_ :str = {
"""val_loss""": val_loss_mean,
"""accuracy_score""": accuracy_score(snake_case , snake_case ),
"""precision""": precision_score(snake_case , snake_case ),
"""recall""": recall_score(snake_case , snake_case ),
"""f1""": fa_score(snake_case , snake_case ),
}
snake_case_ :List[Any] = dict(results.items() )
snake_case_ :Union[str, Any] = results
return ret, preds_list, out_label_list
def lowerCAmelCase_ ( self: Optional[Any] , snake_case: Dict ) -> Optional[Any]:
# when stable
snake_case_, snake_case_, snake_case_ :Tuple = self._eval_end(snake_case )
snake_case_ :str = ret["""log"""]
return {"val_loss": logs["val_loss"], "log": logs, "progress_bar": logs}
def lowerCAmelCase_ ( self: Tuple , snake_case: Optional[int] ) -> Any:
# updating to test_epoch_end instead of deprecated test_end
snake_case_, snake_case_, snake_case_ :Any = self._eval_end(snake_case )
# Converting to the dict required by pl
# https://github.com/PyTorchLightning/pytorch-lightning/blob/master/\
# pytorch_lightning/trainer/logging.py#L139
snake_case_ :Optional[int] = ret["""log"""]
# `val_loss` is the key returned by `self._eval_end()` but actually refers to `test_loss`
return {"avg_test_loss": logs["val_loss"], "log": logs, "progress_bar": logs}
@staticmethod
def lowerCAmelCase_ ( snake_case: Any , snake_case: int ) -> Dict:
# Add NER specific options
BaseTransformer.add_model_specific_args(snake_case , snake_case )
parser.add_argument(
"""--task_type""" , default="""NER""" , type=snake_case , help="""Task type to fine tune in training (e.g. NER, POS, etc)""" )
parser.add_argument(
"""--max_seq_length""" , default=128 , type=snake_case , help=(
"""The maximum total input sequence length after tokenization. Sequences longer """
"""than this will be truncated, sequences shorter will be padded."""
) , )
parser.add_argument(
"""--labels""" , default="""""" , type=snake_case , help="""Path to a file containing all labels. If not specified, CoNLL-2003 labels are used.""" , )
parser.add_argument(
"""--gpus""" , default=0 , type=snake_case , help="""The number of GPUs allocated for this, it is by default 0 meaning none""" , )
parser.add_argument(
"""--overwrite_cache""" , action="""store_true""" , help="""Overwrite the cached training and evaluation sets""" )
return parser
if __name__ == "__main__":
__a = argparse.ArgumentParser()
add_generic_args(parser, os.getcwd())
__a = NERTransformer.add_model_specific_args(parser, os.getcwd())
__a = parser.parse_args()
__a = NERTransformer(args)
__a = generic_train(model, args)
if args.do_predict:
# See https://github.com/huggingface/transformers/issues/3159
# pl use this default format to create a checkpoint:
# https://github.com/PyTorchLightning/pytorch-lightning/blob/master\
# /pytorch_lightning/callbacks/model_checkpoint.py#L322
__a = sorted(glob.glob(os.path.join(args.output_dir, "checkpoint-epoch=*.ckpt"), recursive=True))
__a = model.load_from_checkpoint(checkpoints[-1])
trainer.test(model)
| 66 | 1 |
"""simple docstring"""
import math
class lowerCamelCase :
'''simple docstring'''
def lowerCAmelCase_ ( self: Tuple , snake_case: list[list[float]] , snake_case: list[int] ) -> int:
snake_case_ :Any = 0.0
snake_case_ :Tuple = 0.0
for i in range(len(snake_case ) ):
da += math.pow((sample[i] - weights[0][i]) , 2 )
da += math.pow((sample[i] - weights[1][i]) , 2 )
return 0 if da > da else 1
return 0
def lowerCAmelCase_ ( self: Optional[int] , snake_case: list[list[int | float]] , snake_case: list[int] , snake_case: int , snake_case: float ) -> list[list[int | float]]:
for i in range(len(snake_case ) ):
weights[j][i] += alpha * (sample[i] - weights[j][i])
return weights
def A_ ( ):
'''simple docstring'''
snake_case_ :Dict = [[1, 1, 0, 0], [0, 0, 0, 1], [1, 0, 0, 0], [0, 0, 1, 1]]
# weight initialization ( n, C )
snake_case_ :List[Any] = [[0.2, 0.6, 0.5, 0.9], [0.8, 0.4, 0.7, 0.3]]
# training
snake_case_ :Optional[Any] = SelfOrganizingMap()
snake_case_ :Dict = 3
snake_case_ :Dict = 0.5
for _ in range(_lowercase ):
for j in range(len(_lowercase ) ):
# training sample
snake_case_ :List[Any] = training_samples[j]
# Compute the winning vector
snake_case_ :Optional[int] = self_organizing_map.get_winner(_lowercase, _lowercase )
# Update the winning vector
snake_case_ :List[str] = self_organizing_map.update(_lowercase, _lowercase, _lowercase, _lowercase )
# classify test sample
snake_case_ :str = [0, 0, 0, 1]
snake_case_ :List[Any] = self_organizing_map.get_winner(_lowercase, _lowercase )
# results
print(f"""Clusters that the test sample belongs to : {winner}""" )
print(f"""Weights that have been trained : {weights}""" )
# running the main() function
if __name__ == "__main__":
main()
| 66 |
"""simple docstring"""
from math import factorial
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Optional[int] , snake_case: Dict , snake_case: int ) -> Tuple:
snake_case_ :List[Any] = real
if isinstance(snake_case , snake_case ):
snake_case_ :Tuple = [1] * rank
else:
snake_case_ :Optional[Any] = rank
def __repr__( self: List[str] ) -> Tuple:
return (
f"""{self.real}+"""
f"""{'+'.join(str(snake_case )+'E'+str(n+1 )for n,dual in enumerate(self.duals ) )}"""
)
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
snake_case_ :Any = self.duals.copy()
while cur[-1] == 0:
cur.pop(-1 )
return Dual(self.real , snake_case )
def __add__( self: Optional[int] , snake_case: Dict ) -> List[str]:
if not isinstance(snake_case , snake_case ):
return Dual(self.real + other , self.duals )
snake_case_ :List[Any] = self.duals.copy()
snake_case_ :Tuple = other.duals.copy()
if len(snake_case ) > len(snake_case ):
o_dual.extend([1] * (len(snake_case ) - len(snake_case )) )
elif len(snake_case ) < len(snake_case ):
s_dual.extend([1] * (len(snake_case ) - len(snake_case )) )
snake_case_ :Dict = []
for i in range(len(snake_case ) ):
new_duals.append(s_dual[i] + o_dual[i] )
return Dual(self.real + other.real , snake_case )
_A : str = __add__
def __sub__( self: Tuple , snake_case: Union[str, Any] ) -> Tuple:
return self + other * -1
def __mul__( self: str , snake_case: Tuple ) -> Optional[Any]:
if not isinstance(snake_case , snake_case ):
snake_case_ :Dict = []
for i in self.duals:
new_duals.append(i * other )
return Dual(self.real * other , snake_case )
snake_case_ :int = [0] * (len(self.duals ) + len(other.duals ) + 1)
for i, item in enumerate(self.duals ):
for j, jtem in enumerate(other.duals ):
new_duals[i + j + 1] += item * jtem
for k in range(len(self.duals ) ):
new_duals[k] += self.duals[k] * other.real
for index in range(len(other.duals ) ):
new_duals[index] += other.duals[index] * self.real
return Dual(self.real * other.real , snake_case )
_A : int = __mul__
def __truediv__( self: List[str] , snake_case: List[str] ) -> List[str]:
if not isinstance(snake_case , snake_case ):
snake_case_ :Optional[Any] = []
for i in self.duals:
new_duals.append(i / other )
return Dual(self.real / other , snake_case )
raise ValueError
def __floordiv__( self: int , snake_case: List[Any] ) -> Any:
if not isinstance(snake_case , snake_case ):
snake_case_ :Optional[int] = []
for i in self.duals:
new_duals.append(i // other )
return Dual(self.real // other , snake_case )
raise ValueError
def __pow__( self: Optional[Any] , snake_case: Optional[int] ) -> List[Any]:
if n < 0 or isinstance(snake_case , snake_case ):
raise ValueError("""power must be a positive integer""" )
if n == 0:
return 1
if n == 1:
return self
snake_case_ :str = self
for _ in range(n - 1 ):
x *= self
return x
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
if not callable(_lowercase ):
raise ValueError("""differentiate() requires a function as input for func""" )
if not isinstance(_lowercase, (float, int) ):
raise ValueError("""differentiate() requires a float as input for position""" )
if not isinstance(_lowercase, _lowercase ):
raise ValueError("""differentiate() requires an int as input for order""" )
snake_case_ :Optional[Any] = Dual(_lowercase, 1 )
snake_case_ :List[Any] = func(_lowercase )
if order == 0:
return result.real
return result.duals[order - 1] * factorial(_lowercase )
if __name__ == "__main__":
import doctest
doctest.testmod()
def A_ ( _lowercase ):
'''simple docstring'''
return y**2 * y**4
print(differentiate(f, 9, 2))
| 66 | 1 |
"""simple docstring"""
import os
import unicodedata
from shutil import copyfile
from typing import Any, Dict, List, Optional, Tuple
import sentencepiece as spm
from ...tokenization_utils import AddedToken, PreTrainedTokenizer
from ...utils import SPIECE_UNDERLINE, logging
__a = logging.get_logger(__name__)
__a = {"vocab_file": "spiece.model"}
__a = {
"vocab_file": {
"TsinghuaAI/CPM-Generate": "https://huggingface.co/TsinghuaAI/CPM-Generate/resolve/main/spiece.model",
}
}
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
def __init__( self: Tuple , snake_case: Any , snake_case: Any=False , snake_case: Any=True , snake_case: List[Any]=False , snake_case: Tuple="<s>" , snake_case: List[str]="</s>" , snake_case: Tuple="<unk>" , snake_case: Optional[Any]="<sep>" , snake_case: Any="<pad>" , snake_case: Optional[Any]="<cls>" , snake_case: int="<mask>" , snake_case: List[Any]=["<eop>", "<eod>"] , snake_case: Optional[Dict[str, Any]] = None , **snake_case: int , ) -> None:
snake_case_ :Dict = AddedToken(snake_case , lstrip=snake_case , rstrip=snake_case ) if isinstance(snake_case , snake_case ) else mask_token
snake_case_ :int = {} if sp_model_kwargs is None else sp_model_kwargs
super().__init__(
do_lower_case=snake_case , remove_space=snake_case , keep_accents=snake_case , bos_token=snake_case , eos_token=snake_case , unk_token=snake_case , sep_token=snake_case , pad_token=snake_case , cls_token=snake_case , mask_token=snake_case , additional_special_tokens=snake_case , sp_model_kwargs=self.sp_model_kwargs , **snake_case , )
snake_case_ :str = 3
snake_case_ :List[Any] = do_lower_case
snake_case_ :List[str] = remove_space
snake_case_ :str = keep_accents
snake_case_ :str = vocab_file
snake_case_ :Any = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(snake_case )
try:
import jieba
except ModuleNotFoundError as error:
raise error.__class__(
"""You need to install jieba to use CpmTokenizer or CpmTokenizerFast. """
"""See https://pypi.org/project/jieba/ for installation.""" )
snake_case_ :List[str] = jieba
snake_case_ :str = str.maketrans(""" \n""" , """\u2582\u2583""" )
@property
# Copied from transformers.models.xlnet.tokenization_xlnet.XLNetTokenizer.vocab_size
def lowerCAmelCase_ ( self: Dict ) -> Tuple:
return len(self.sp_model )
def lowerCAmelCase_ ( self: int ) -> Any:
snake_case_ :Any = {self.convert_ids_to_tokens(snake_case ): i for i in range(self.vocab_size )}
vocab.update(self.added_tokens_encoder )
return vocab
def __getstate__( self: Dict ) -> Dict:
snake_case_ :Optional[Any] = self.__dict__.copy()
snake_case_ :Dict = None
return state
def __setstate__( self: int , snake_case: str ) -> Tuple:
snake_case_ :Tuple = d
# for backward compatibility
if not hasattr(self , """sp_model_kwargs""" ):
snake_case_ :int = {}
snake_case_ :Optional[Any] = spm.SentencePieceProcessor(**self.sp_model_kwargs )
self.sp_model.Load(self.vocab_file )
def lowerCAmelCase_ ( self: Any , snake_case: List[Any] ) -> str:
if self.remove_space:
snake_case_ :Dict = """ """.join(inputs.strip().split() )
else:
snake_case_ :Any = inputs
snake_case_ :Optional[Any] = outputs.replace("""``""" , """\"""" ).replace("""''""" , """\"""" )
if not self.keep_accents:
snake_case_ :Any = unicodedata.normalize("""NFKD""" , snake_case )
snake_case_ :int = """""".join([c for c in outputs if not unicodedata.combining(snake_case )] )
if self.do_lower_case:
snake_case_ :Tuple = outputs.lower()
return outputs
def lowerCAmelCase_ ( self: Optional[int] , snake_case: str ) -> List[str]:
snake_case_ :str = self.preprocess_text(snake_case )
snake_case_ :Optional[Any] = self.sp_model.encode(snake_case , out_type=snake_case )
snake_case_ :int = []
for piece in pieces:
if len(snake_case ) > 1 and piece[-1] == str(""",""" ) and piece[-2].isdigit():
snake_case_ :List[str] = self.sp_model.EncodeAsPieces(piece[:-1].replace(snake_case , """""" ) )
if piece[0] != SPIECE_UNDERLINE and cur_pieces[0][0] == SPIECE_UNDERLINE:
if len(cur_pieces[0] ) == 1:
snake_case_ :Any = cur_pieces[1:]
else:
snake_case_ :Optional[int] = cur_pieces[0][1:]
cur_pieces.append(piece[-1] )
new_pieces.extend(snake_case )
else:
new_pieces.append(snake_case )
return new_pieces
def lowerCAmelCase_ ( self: int , snake_case: str ) -> Dict:
return self.sp_model.PieceToId(snake_case )
def lowerCAmelCase_ ( self: Dict , snake_case: int ) -> List[str]:
return self.sp_model.IdToPiece(snake_case )
def lowerCAmelCase_ ( self: Union[str, Any] , snake_case: Union[str, Any] ) -> List[Any]:
snake_case_ :str = """""".join(snake_case ).replace(snake_case , """ """ ).strip()
return out_string
def lowerCAmelCase_ ( self: int , snake_case: List[int] , snake_case: Optional[List[int]] = None ) -> List[int]:
snake_case_ :Tuple = [self.sep_token_id]
snake_case_ :Tuple = [self.cls_token_id]
if token_ids_a is None:
return token_ids_a + sep + cls
return token_ids_a + sep + token_ids_a + sep + cls
def lowerCAmelCase_ ( self: int , snake_case: List[int] , snake_case: Optional[List[int]] = None , snake_case: bool = False ) -> List[int]:
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 )
if token_ids_a is not None:
return ([0] * len(snake_case )) + [1] + ([0] * len(snake_case )) + [1, 1]
return ([0] * len(snake_case )) + [1, 1]
def lowerCAmelCase_ ( self: int , snake_case: List[int] , snake_case: Optional[List[int]] = None ) -> List[int]:
snake_case_ :List[Any] = [self.sep_token_id]
snake_case_ :str = [2]
if token_ids_a is None:
return len(token_ids_a + sep ) * [0] + cls_segment_id
return len(token_ids_a + sep ) * [0] + len(token_ids_a + sep ) * [1] + cls_segment_id
def lowerCAmelCase_ ( self: Union[str, Any] , snake_case: str , snake_case: Optional[str] = None ) -> Tuple[str]:
if not os.path.isdir(snake_case ):
logger.error(f"""Vocabulary path ({save_directory}) should be a directory""" )
return
snake_case_ :List[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_ :Union[str, Any] = self.sp_model.serialized_model_proto()
fi.write(snake_case )
return (out_vocab_file,)
def lowerCAmelCase_ ( self: List[str] , *snake_case: str , **snake_case: List[Any] ) -> Tuple:
snake_case_ :int = super()._decode(*snake_case , **snake_case )
snake_case_ :Optional[int] = text.replace(""" """ , """""" ).replace("""\u2582""" , """ """ ).replace("""\u2583""" , """\n""" )
return text
| 66 |
"""simple docstring"""
from __future__ import annotations
__a = 10
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Union[str, Any] = 1
snake_case_ :List[str] = max(_lowercase )
while placement <= max_digit:
# declare and initialize empty buckets
snake_case_ :list[list] = [[] for _ in range(_lowercase )]
# split list_of_ints between the buckets
for i in list_of_ints:
snake_case_ :Any = int((i / placement) % RADIX )
buckets[tmp].append(_lowercase )
# put each buckets' contents into list_of_ints
snake_case_ :Optional[Any] = 0
for b in range(_lowercase ):
for i in buckets[b]:
snake_case_ :Union[str, Any] = i
a += 1
# move to next
placement *= RADIX
return list_of_ints
if __name__ == "__main__":
import doctest
doctest.testmod()
| 66 | 1 |
"""simple docstring"""
import unittest
import numpy as np
import torch
from diffusers import PNDMPipeline, PNDMScheduler, UNetaDModel
from diffusers.utils.testing_utils import enable_full_determinism, require_torch, slow, torch_device
enable_full_determinism()
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
@property
def lowerCAmelCase_ ( self: str ) -> Union[str, Any]:
torch.manual_seed(0 )
snake_case_ :str = UNetaDModel(
block_out_channels=(32, 64) , layers_per_block=2 , sample_size=32 , in_channels=3 , out_channels=3 , down_block_types=("""DownBlock2D""", """AttnDownBlock2D""") , up_block_types=("""AttnUpBlock2D""", """UpBlock2D""") , )
return model
def lowerCAmelCase_ ( self: List[str] ) -> str:
snake_case_ :Union[str, Any] = self.dummy_uncond_unet
snake_case_ :Tuple = PNDMScheduler()
snake_case_ :str = PNDMPipeline(unet=snake_case , scheduler=snake_case )
pndm.to(snake_case )
pndm.set_progress_bar_config(disable=snake_case )
snake_case_ :int = torch.manual_seed(0 )
snake_case_ :List[Any] = pndm(generator=snake_case , num_inference_steps=20 , output_type="""numpy""" ).images
snake_case_ :Optional[Any] = torch.manual_seed(0 )
snake_case_ :Union[str, Any] = pndm(generator=snake_case , num_inference_steps=20 , output_type="""numpy""" , return_dict=snake_case )[0]
snake_case_ :Optional[int] = image[0, -3:, -3:, -1]
snake_case_ :str = image_from_tuple[0, -3:, -3:, -1]
assert image.shape == (1, 32, 32, 3)
snake_case_ :List[str] = np.array([1.0, 1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2
assert np.abs(image_from_tuple_slice.flatten() - expected_slice ).max() < 1E-2
@slow
@require_torch
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: List[Any] ) -> Union[str, Any]:
snake_case_ :int = """google/ddpm-cifar10-32"""
snake_case_ :List[str] = UNetaDModel.from_pretrained(snake_case )
snake_case_ :Any = PNDMScheduler()
snake_case_ :Tuple = PNDMPipeline(unet=snake_case , scheduler=snake_case )
pndm.to(snake_case )
pndm.set_progress_bar_config(disable=snake_case )
snake_case_ :str = torch.manual_seed(0 )
snake_case_ :Any = pndm(generator=snake_case , output_type="""numpy""" ).images
snake_case_ :Optional[int] = image[0, -3:, -3:, -1]
assert image.shape == (1, 32, 32, 3)
snake_case_ :Optional[int] = np.array([0.1_5_6_4, 0.1_4_6_4_5, 0.1_4_0_6, 0.1_4_7_1_5, 0.1_2_4_2_5, 0.1_4_0_4_5, 0.1_3_1_1_5, 0.1_2_1_7_5, 0.1_2_5] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 1E-2
| 66 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_sentencepiece_available,
is_tokenizers_available,
is_torch_available,
)
__a = {"configuration_reformer": ["REFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP", "ReformerConfig"]}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["ReformerTokenizer"]
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["ReformerTokenizerFast"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"REFORMER_PRETRAINED_MODEL_ARCHIVE_LIST",
"ReformerAttention",
"ReformerForMaskedLM",
"ReformerForQuestionAnswering",
"ReformerForSequenceClassification",
"ReformerLayer",
"ReformerModel",
"ReformerModelWithLMHead",
"ReformerPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_reformer import REFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP, ReformerConfig
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_reformer import ReformerTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_reformer_fast import ReformerTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_reformer import (
REFORMER_PRETRAINED_MODEL_ARCHIVE_LIST,
ReformerAttention,
ReformerForMaskedLM,
ReformerForQuestionAnswering,
ReformerForSequenceClassification,
ReformerLayer,
ReformerModel,
ReformerModelWithLMHead,
ReformerPreTrainedModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 | 1 |
"""simple docstring"""
def A_ ( _lowercase = 1000000 ):
'''simple docstring'''
snake_case_ :Union[str, Any] = 1
snake_case_ :Optional[int] = 1
snake_case_ :Optional[int] = {1: 1}
for inputa in range(2, _lowercase ):
snake_case_ :List[str] = 0
snake_case_ :Optional[int] = inputa
while True:
if number in counters:
counter += counters[number]
break
if number % 2 == 0:
number //= 2
counter += 1
else:
snake_case_ :int = (3 * number) + 1
counter += 1
if inputa not in counters:
snake_case_ :Optional[Any] = counter
if counter > pre_counter:
snake_case_ :str = inputa
snake_case_ :List[str] = counter
return largest_number
if __name__ == "__main__":
print(solution(int(input().strip())))
| 66 |
"""simple docstring"""
import gc
import unittest
from diffusers import FlaxControlNetModel, FlaxStableDiffusionControlNetPipeline
from diffusers.utils import is_flax_available, load_image, slow
from diffusers.utils.testing_utils import require_flax
if is_flax_available():
import jax
import jax.numpy as jnp
from flax.jax_utils import replicate
from flax.training.common_utils import shard
@slow
@require_flax
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: List[Any] ) -> Any:
# clean up the VRAM after each test
super().tearDown()
gc.collect()
def lowerCAmelCase_ ( self: Tuple ) -> Any:
snake_case_, snake_case_ :List[str] = FlaxControlNetModel.from_pretrained(
"""lllyasviel/sd-controlnet-canny""" , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_, snake_case_ :Union[str, Any] = FlaxStableDiffusionControlNetPipeline.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" , controlnet=snake_case , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_ :Union[str, Any] = controlnet_params
snake_case_ :Union[str, Any] = """bird"""
snake_case_ :List[Any] = jax.device_count()
snake_case_ :List[Any] = pipe.prepare_text_inputs([prompts] * num_samples )
snake_case_ :List[str] = load_image(
"""https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/sd_controlnet/bird_canny.png""" )
snake_case_ :List[str] = pipe.prepare_image_inputs([canny_image] * num_samples )
snake_case_ :Any = jax.random.PRNGKey(0 )
snake_case_ :List[str] = jax.random.split(snake_case , jax.device_count() )
snake_case_ :List[Any] = replicate(snake_case )
snake_case_ :List[str] = shard(snake_case )
snake_case_ :str = shard(snake_case )
snake_case_ :Dict = pipe(
prompt_ids=snake_case , image=snake_case , params=snake_case , prng_seed=snake_case , num_inference_steps=50 , jit=snake_case , ).images
assert images.shape == (jax.device_count(), 1, 768, 512, 3)
snake_case_ :str = images.reshape((images.shape[0] * images.shape[1],) + images.shape[-3:] )
snake_case_ :Union[str, Any] = images[0, 253:256, 253:256, -1]
snake_case_ :str = jnp.asarray(jax.device_get(image_slice.flatten() ) )
snake_case_ :Dict = jnp.array(
[0.1_6_7_9_6_9, 0.1_1_6_6_9_9, 0.0_8_1_5_4_3, 0.1_5_4_2_9_7, 0.1_3_2_8_1_2, 0.1_0_8_8_8_7, 0.1_6_9_9_2_2, 0.1_6_9_9_2_2, 0.2_0_5_0_7_8] )
print(f"""output_slice: {output_slice}""" )
assert jnp.abs(output_slice - expected_slice ).max() < 1E-2
def lowerCAmelCase_ ( self: int ) -> Dict:
snake_case_, snake_case_ :List[Any] = FlaxControlNetModel.from_pretrained(
"""lllyasviel/sd-controlnet-openpose""" , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_, snake_case_ :int = FlaxStableDiffusionControlNetPipeline.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" , controlnet=snake_case , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_ :str = controlnet_params
snake_case_ :Optional[int] = """Chef in the kitchen"""
snake_case_ :Union[str, Any] = jax.device_count()
snake_case_ :Any = pipe.prepare_text_inputs([prompts] * num_samples )
snake_case_ :str = load_image(
"""https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/sd_controlnet/pose.png""" )
snake_case_ :Optional[Any] = pipe.prepare_image_inputs([pose_image] * num_samples )
snake_case_ :str = jax.random.PRNGKey(0 )
snake_case_ :str = jax.random.split(snake_case , jax.device_count() )
snake_case_ :Tuple = replicate(snake_case )
snake_case_ :str = shard(snake_case )
snake_case_ :int = shard(snake_case )
snake_case_ :List[str] = pipe(
prompt_ids=snake_case , image=snake_case , params=snake_case , prng_seed=snake_case , num_inference_steps=50 , jit=snake_case , ).images
assert images.shape == (jax.device_count(), 1, 768, 512, 3)
snake_case_ :str = images.reshape((images.shape[0] * images.shape[1],) + images.shape[-3:] )
snake_case_ :int = images[0, 253:256, 253:256, -1]
snake_case_ :Dict = jnp.asarray(jax.device_get(image_slice.flatten() ) )
snake_case_ :Optional[int] = jnp.array(
[[0.2_7_1_4_8_4, 0.2_6_1_7_1_9, 0.2_7_5_3_9_1, 0.2_7_7_3_4_4, 0.2_7_9_2_9_7, 0.2_9_1_0_1_6, 0.2_9_4_9_2_2, 0.3_0_2_7_3_4, 0.3_0_2_7_3_4]] )
print(f"""output_slice: {output_slice}""" )
assert jnp.abs(output_slice - expected_slice ).max() < 1E-2
| 66 | 1 |
"""simple docstring"""
__a = "\n# Transformers installation\n! pip install transformers datasets\n# To install from source instead of the last release, comment the command above and uncomment the following one.\n# ! pip install git+https://github.com/huggingface/transformers.git\n"
__a = [{"type": "code", "content": INSTALL_CONTENT}]
__a = {
"{processor_class}": "FakeProcessorClass",
"{model_class}": "FakeModelClass",
"{object_class}": "FakeObjectClass",
}
| 66 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
__a = {
"configuration_mobilebert": [
"MOBILEBERT_PRETRAINED_CONFIG_ARCHIVE_MAP",
"MobileBertConfig",
"MobileBertOnnxConfig",
],
"tokenization_mobilebert": ["MobileBertTokenizer"],
}
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["MobileBertTokenizerFast"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST",
"MobileBertForMaskedLM",
"MobileBertForMultipleChoice",
"MobileBertForNextSentencePrediction",
"MobileBertForPreTraining",
"MobileBertForQuestionAnswering",
"MobileBertForSequenceClassification",
"MobileBertForTokenClassification",
"MobileBertLayer",
"MobileBertModel",
"MobileBertPreTrainedModel",
"load_tf_weights_in_mobilebert",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"TF_MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST",
"TFMobileBertForMaskedLM",
"TFMobileBertForMultipleChoice",
"TFMobileBertForNextSentencePrediction",
"TFMobileBertForPreTraining",
"TFMobileBertForQuestionAnswering",
"TFMobileBertForSequenceClassification",
"TFMobileBertForTokenClassification",
"TFMobileBertMainLayer",
"TFMobileBertModel",
"TFMobileBertPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_mobilebert import (
MOBILEBERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
MobileBertConfig,
MobileBertOnnxConfig,
)
from .tokenization_mobilebert import MobileBertTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_mobilebert_fast import MobileBertTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_mobilebert import (
MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST,
MobileBertForMaskedLM,
MobileBertForMultipleChoice,
MobileBertForNextSentencePrediction,
MobileBertForPreTraining,
MobileBertForQuestionAnswering,
MobileBertForSequenceClassification,
MobileBertForTokenClassification,
MobileBertLayer,
MobileBertModel,
MobileBertPreTrainedModel,
load_tf_weights_in_mobilebert,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_mobilebert import (
TF_MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST,
TFMobileBertForMaskedLM,
TFMobileBertForMultipleChoice,
TFMobileBertForNextSentencePrediction,
TFMobileBertForPreTraining,
TFMobileBertForQuestionAnswering,
TFMobileBertForSequenceClassification,
TFMobileBertForTokenClassification,
TFMobileBertMainLayer,
TFMobileBertModel,
TFMobileBertPreTrainedModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 | 1 |
"""simple docstring"""
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :Optional[int] = word.split()
def justify(_lowercase, _lowercase, _lowercase ) -> str:
snake_case_ :Union[str, Any] = max_width - width
snake_case_ :Dict = len(_lowercase )
if len(_lowercase ) == 1:
# if there is only word in line
# just insert overall_spaces_count for the remainder of line
return line[0] + " " * overall_spaces_count
else:
snake_case_ :Optional[Any] = words_count - 1
# num_spaces_between_words_list[i] : tells you to insert
# num_spaces_between_words_list[i] spaces
# after word on line[i]
snake_case_ :Union[str, Any] = spaces_to_insert_between_words * [
overall_spaces_count // spaces_to_insert_between_words
]
snake_case_ :Tuple = (
overall_spaces_count % spaces_to_insert_between_words
)
# distribute spaces via round robin to the left words
for i in range(_lowercase ):
num_spaces_between_words_list[i] += 1
snake_case_ :Union[str, Any] = []
for i in range(_lowercase ):
# add the word
aligned_words_list.append(line[i] )
# add the spaces to insert
aligned_words_list.append(num_spaces_between_words_list[i] * """ """ )
# just add the last word to the sentence
aligned_words_list.append(line[-1] )
# join the aligned words list to form a justified line
return "".join(_lowercase )
snake_case_ :Union[str, Any] = []
snake_case_ :list[str] = []
snake_case_ :Optional[Any] = 0
for word in words:
if width + len(_lowercase ) + len(_lowercase ) <= max_width:
# keep adding words until we can fill out max_width
# width = sum of length of all words (without overall_spaces_count)
# len(word) = length of current word
# len(line) = number of overall_spaces_count to insert between words
line.append(_lowercase )
width += len(_lowercase )
else:
# justify the line and add it to result
answer.append(justify(_lowercase, _lowercase, _lowercase ) )
# reset new line and new width
snake_case_, snake_case_ :Tuple = [word], len(_lowercase )
snake_case_ :Any = max_width - width - len(_lowercase )
answer.append(""" """.join(_lowercase ) + (remaining_spaces + 1) * """ """ )
return answer
if __name__ == "__main__":
from doctest import testmod
testmod()
| 66 |
"""simple docstring"""
import argparse
import json
import os
from collections import OrderedDict
import numpy as np
import tensorflow as tf
import torch
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Union[str, Any] = os.path.join(args.tf_model_dir, """parameters.json""" )
snake_case_ :Any = json.loads(open(_lowercase ).read() )
if not params:
raise ValueError(
f"""It seems that the json file at {parameter_file} is empty. Make sure you have a correct json file.""" )
if not args.output.endswith(""".pt""" ):
snake_case_ :Optional[int] = args.output + """.pt"""
snake_case_ :List[str] = OrderedDict()
with tf.device("""/CPU:0""" ):
snake_case_ :Dict = tf.train.load_checkpoint(args.tf_model_dir )
snake_case_ :str = reader.get_variable_to_shape_map()
for key_name in shapes.keys():
snake_case_ :List[Any] = reader.get_tensor(_lowercase ).astype(np.floataa )
if key_name.endswith("""/adam_m""" ) or key_name.endswith("""/adam_v""" ):
continue
if key_name.startswith("""pasts/""" ):
if key_name.startswith("""pasts/mlp""" ):
snake_case_ :Any = int(key_name[9] )
elif key_name.startswith("""pasts/out""" ):
snake_case_ :Optional[int] = 8
snake_case_ :List[str] = """model.sqout.%d.weight""" % (player * 2) # enter to nn.Sequencial with Tanh, so 2 at a time
snake_case_ :Optional[Any] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :List[str] = torch.tensor(_lowercase )
elif key_name.startswith("""model/moe""" ):
snake_case_ :Tuple = int(key_name[9:].split("""/""" )[0] )
if key_name.endswith("""/switch_gating/kernel""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.feed_forward.mlp.router.classifier.weight""" % player
snake_case_ :Optional[Any] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/softmlp/kernel""" ):
snake_case_ :List[Any] = """model.blocks.%d.feed_forward.soft_bypass_mlp.weight""" % player
snake_case_ :Optional[int] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/wo/kernel""" ) or key_name.endswith("""/wi/kernel""" ):
snake_case_ :Dict = key_name[-9:-7]
for i in range(16 ):
snake_case_ :str = """model.blocks.%d.feed_forward.mlp.experts.expert_%d.%s.weight""" % (player, i, nlayer)
snake_case_ :Tuple = (
vnp[i].transpose([1, 0] ).copy()
) # In Mesh-Tensorflow, it is one array, so it is divided
snake_case_ :Optional[int] = torch.tensor(_lowercase )
elif key_name.startswith("""model/mlp""" ):
snake_case_ :Optional[int] = int(key_name[9:].split("""/""" )[0] )
if key_name.endswith("""/p1/kernel""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.feed_forward.mlp.wi.weight""" % player
snake_case_ :Dict = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/p1/bias""" ):
snake_case_ :List[Any] = """model.blocks.%d.feed_forward.mlp.wi.bias""" % player
snake_case_ :str = vnp.copy() # same because it is one dimensional
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/p2/kernel""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.feed_forward.mlp.wo.weight""" % player
snake_case_ :Dict = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Tuple = torch.tensor(_lowercase )
elif key_name.endswith("""/p2/bias""" ):
snake_case_ :Dict = """model.blocks.%d.feed_forward.mlp.wo.bias""" % player
snake_case_ :Any = vnp.copy() # same because it is one dimensional
snake_case_ :Optional[int] = torch.tensor(_lowercase )
elif key_name.startswith("""model/ln""" ):
snake_case_ :Union[str, Any] = int(key_name[8:].split("""/""" )[0] )
if key_name.endswith("""/b""" ):
snake_case_ :str = """model.blocks.%d.feed_forward.norm.bias""" % player
snake_case_ :Dict = vnp.copy() # same because it is one dimensional
snake_case_ :int = torch.tensor(_lowercase )
elif key_name.endswith("""/g""" ):
snake_case_ :Dict = """model.blocks.%d.feed_forward.norm.weight""" % player
snake_case_ :Dict = vnp.copy() # same because it is one dimensional
snake_case_ :Tuple = torch.tensor(_lowercase )
elif key_name.startswith("""model/att""" ):
snake_case_ :List[str] = int(key_name[9:].split("""/""" )[0] )
if key_name.endswith("""/qkv/kernel""" ):
snake_case_ :Optional[int] = vnp.copy() # Compute same dimension as Mesh-tensorflow using einsum
snake_case_ :Dict = state[:, 0, :, :]
snake_case_ :int = state[:, 1, :, :]
snake_case_ :List[str] = state[:, 2, :, :]
snake_case_ :str = (
state_q.reshape([state_q.shape[0], state_q.shape[1] * state_q.shape[2]] )
.transpose([1, 0] )
.copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Any = (
state_k.reshape([state_k.shape[0], state_k.shape[1] * state_k.shape[2]] )
.transpose([1, 0] )
.copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[int] = (
state_v.reshape([state_v.shape[0], state_v.shape[1] * state_v.shape[2]] )
.transpose([1, 0] )
.copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :int = """model.blocks.%d.self_attn.self_attn.q_proj.weight""" % player
snake_case_ :int = torch.tensor(_lowercase )
snake_case_ :Optional[Any] = """model.blocks.%d.self_attn.self_attn.k_proj.weight""" % player
snake_case_ :Dict = torch.tensor(_lowercase )
snake_case_ :Dict = """model.blocks.%d.self_attn.self_attn.v_proj.weight""" % player
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/o/kernel""" ):
snake_case_ :str = """model.blocks.%d.self_attn.self_attn.out_proj.weight""" % player
snake_case_ :str = (
vnp.reshape([vnp.shape[0] * vnp.shape[1], vnp.shape[2]] ).transpose([1, 0] ).copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Any = torch.tensor(_lowercase )
elif key_name.startswith("""model/an""" ):
snake_case_ :Optional[int] = int(key_name[8:].split("""/""" )[0] )
if key_name.endswith("""/b""" ):
snake_case_ :Any = """model.blocks.%d.self_attn.norm.bias""" % player
snake_case_ :Optional[int] = vnp.copy() # same because it is one dimensional
snake_case_ :Tuple = torch.tensor(_lowercase )
elif key_name.endswith("""/g""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.self_attn.norm.weight""" % player
snake_case_ :Dict = vnp.copy() # same because it is one dimensional
snake_case_ :Optional[int] = torch.tensor(_lowercase )
elif (
key_name.startswith("""model/wte""" )
or key_name.startswith("""model/wpe""" )
or key_name.startswith("""model/ete""" )
):
snake_case_ :List[Any] = {"""wte""": """embed_tokens""", """wpe""": """position_embeddings""", """ete""": """extra_position_embeddings"""}[
key_name[-3:]
]
snake_case_ :Optional[Any] = """model.%s.weight""" % nlayer
snake_case_ :Any = vnp.copy() # same in embedded
snake_case_ :List[Any] = torch.tensor(_lowercase )
if key_name.startswith("""model/wte""" ):
snake_case_ :Tuple = """lm_head.weight"""
snake_case_ :List[str] = vnp.copy() # same in embedded
snake_case_ :List[Any] = torch.tensor(_lowercase )
elif key_name.startswith("""model/wob""" ):
snake_case_ :str = """final_logits_bias"""
snake_case_ :Any = vnp.copy() # same in embedded
snake_case_ :List[Any] = state.reshape((1, -1) )
snake_case_ :Union[str, Any] = torch.tensor(_lowercase )
elif key_name == "model/dense/kernel":
snake_case_ :str = """model.last_project.weight"""
snake_case_ :Dict = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :int = torch.tensor(_lowercase )
elif key_name == "model/dense_1/bias":
snake_case_ :Optional[int] = """model.last_project.bias"""
snake_case_ :Tuple = vnp.copy() # same because it is one dimensional
snake_case_ :Any = torch.tensor(_lowercase )
torch.save(_lowercase, args.output )
if __name__ == "__main__":
__a = argparse.ArgumentParser(
description="model converter.", formatter_class=argparse.ArgumentDefaultsHelpFormatter
)
parser.add_argument("--tf_model_dir", metavar="PATH", type=str, required=True, help="import model")
parser.add_argument("--output", metavar="PATH", type=str, required=True, help="output model")
__a = parser.parse_args()
convert_tf_gptsan_to_pt(args)
| 66 | 1 |
"""simple docstring"""
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ...tokenization_utils_fast import PreTrainedTokenizerFast
from .tokenization_lxmert import LxmertTokenizer
__a = {"vocab_file": "vocab.txt", "tokenizer_file": "tokenizer.json"}
__a = {
"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"
),
},
}
__a = {
"unc-nlp/lxmert-base-uncased": 5_12,
}
__a = {
"unc-nlp/lxmert-base-uncased": {"do_lower_case": True},
}
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
_A : str = VOCAB_FILES_NAMES
_A : int = PRETRAINED_VOCAB_FILES_MAP
_A : Dict = PRETRAINED_INIT_CONFIGURATION
_A : List[str] = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
_A : Dict = LxmertTokenizer
def __init__( self: str , snake_case: Optional[int]=None , snake_case: Dict=None , snake_case: Union[str, Any]=True , snake_case: str="[UNK]" , snake_case: List[str]="[SEP]" , snake_case: Dict="[PAD]" , snake_case: Dict="[CLS]" , snake_case: Union[str, Any]="[MASK]" , snake_case: Union[str, Any]=True , snake_case: Tuple=None , **snake_case: Optional[Any] , ) -> Tuple:
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_ :Union[str, Any] = 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_ :List[Any] = getattr(snake_case , normalizer_state.pop("""type""" ) )
snake_case_ :Union[str, Any] = do_lower_case
snake_case_ :Optional[int] = strip_accents
snake_case_ :int = tokenize_chinese_chars
snake_case_ :str = normalizer_class(**snake_case )
snake_case_ :List[str] = do_lower_case
def lowerCAmelCase_ ( self: str , snake_case: Optional[Any] , snake_case: Union[str, Any]=None ) -> List[Any]:
snake_case_ :Dict = [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 lowerCAmelCase_ ( self: List[str] , snake_case: List[int] , snake_case: Optional[List[int]] = None ) -> List[int]:
snake_case_ :int = [self.sep_token_id]
snake_case_ :List[str] = [self.cls_token_id]
if token_ids_a is None:
return len(cls + token_ids_a + sep ) * [0]
return len(cls + token_ids_a + sep ) * [0] + len(token_ids_a + sep ) * [1]
def lowerCAmelCase_ ( self: Dict , snake_case: str , snake_case: Optional[str] = None ) -> Tuple[str]:
snake_case_ :List[str] = self._tokenizer.model.save(snake_case , name=snake_case )
return tuple(snake_case )
| 66 |
"""simple docstring"""
import numpy as np
import pandas as pd
from sklearn.preprocessing import MinMaxScaler
from tensorflow.keras.layers import LSTM, Dense
from tensorflow.keras.models import Sequential
if __name__ == "__main__":
__a = pd.read_csv("sample_data.csv", header=None)
__a = df.shape[:1][0]
# If you're using some other dataset input the target column
__a = df.iloc[:, 1:2]
__a = actual_data.values.reshape(len_data, 1)
__a = MinMaxScaler().fit_transform(actual_data)
__a = 10
__a = 5
__a = 20
__a = len_data - periods * look_back
__a = actual_data[:division]
__a = actual_data[division - look_back :]
__a , __a = [], []
__a , __a = [], []
for i in range(0, len(train_data) - forward_days - look_back + 1):
train_x.append(train_data[i : i + look_back])
train_y.append(train_data[i + look_back : i + look_back + forward_days])
for i in range(0, len(test_data) - forward_days - look_back + 1):
test_x.append(test_data[i : i + look_back])
test_y.append(test_data[i + look_back : i + look_back + forward_days])
__a = np.array(train_x)
__a = np.array(test_x)
__a = np.array([list(i.ravel()) for i in train_y])
__a = np.array([list(i.ravel()) for i in test_y])
__a = Sequential()
model.add(LSTM(1_28, input_shape=(look_back, 1), return_sequences=True))
model.add(LSTM(64, input_shape=(1_28, 1)))
model.add(Dense(forward_days))
model.compile(loss="mean_squared_error", optimizer="adam")
__a = model.fit(
x_train, y_train, epochs=1_50, verbose=1, shuffle=True, batch_size=4
)
__a = model.predict(x_test)
| 66 | 1 |
"""simple docstring"""
import json
import os
import subprocess
import unittest
from ast import literal_eval
import pytest
from parameterized import parameterized, parameterized_class
from . import is_sagemaker_available
if is_sagemaker_available():
from sagemaker import Session, TrainingJobAnalytics
from sagemaker.huggingface import HuggingFace
@pytest.mark.skipif(
literal_eval(os.getenv("""TEST_SAGEMAKER""" , """False""" ) ) is not True , reason="""Skipping test because should only be run when releasing minor transformers version""" , )
@pytest.mark.usefixtures("""sm_env""" )
@parameterized_class(
[
{
"""framework""": """pytorch""",
"""script""": """run_glue_model_parallelism.py""",
"""model_name_or_path""": """roberta-large""",
"""instance_type""": """ml.p3dn.24xlarge""",
"""results""": {"""train_runtime""": 1_6_0_0, """eval_accuracy""": 0.3, """eval_loss""": 1.2},
},
{
"""framework""": """pytorch""",
"""script""": """run_glue.py""",
"""model_name_or_path""": """roberta-large""",
"""instance_type""": """ml.p3dn.24xlarge""",
"""results""": {"""train_runtime""": 1_6_0_0, """eval_accuracy""": 0.3, """eval_loss""": 1.2},
},
] )
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: Any ) -> str:
if self.framework == "pytorch":
subprocess.run(
f"""cp ./examples/pytorch/text-classification/run_glue.py {self.env.test_path}/run_glue.py""".split() , encoding="""utf-8""" , check=snake_case , )
assert hasattr(self , """env""" )
def lowerCAmelCase_ ( self: int , snake_case: Dict ) -> List[Any]:
# configuration for running training on smdistributed Model Parallel
snake_case_ :Tuple = {
"""enabled""": True,
"""processes_per_host""": 8,
}
snake_case_ :List[Any] = {
"""enabled""": True,
"""parameters""": {
"""microbatches""": 4,
"""placement_strategy""": """spread""",
"""pipeline""": """interleaved""",
"""optimize""": """speed""",
"""partitions""": 4,
"""ddp""": True,
},
}
snake_case_ :Tuple = {"""smdistributed""": {"""modelparallel""": smp_options}, """mpi""": mpi_options}
snake_case_ :Any = """trainer""" if self.script == """run_glue.py""" else """smtrainer"""
# creates estimator
return HuggingFace(
entry_point=self.script , source_dir=self.env.test_path , role=self.env.role , image_uri=self.env.image_uri , base_job_name=f"""{self.env.base_job_name}-{instance_count}-smp-{name_extension}""" , instance_count=snake_case , instance_type=self.instance_type , debugger_hook_config=snake_case , hyperparameters={
**self.env.hyperparameters,
"""model_name_or_path""": self.model_name_or_path,
"""max_steps""": 500,
} , metric_definitions=self.env.metric_definitions , distribution=snake_case , py_version="""py36""" , )
def lowerCAmelCase_ ( self: Any , snake_case: Tuple ) -> List[str]:
TrainingJobAnalytics(snake_case ).export_csv(f"""{self.env.test_path}/{job_name}_metrics.csv""" )
@parameterized.expand([(1,)] )
def lowerCAmelCase_ ( self: Dict , snake_case: Dict ) -> List[Any]:
# create estimator
snake_case_ :List[Any] = self.create_estimator(snake_case )
# run training
estimator.fit()
# result dataframe
snake_case_ :Any = TrainingJobAnalytics(estimator.latest_training_job.name ).dataframe()
# extract kpis
snake_case_ :Tuple = list(result_metrics_df[result_metrics_df.metric_name == """eval_accuracy"""]["""value"""] )
snake_case_ :Dict = list(result_metrics_df[result_metrics_df.metric_name == """eval_loss"""]["""value"""] )
# get train time from SageMaker job, this includes starting, preprocessing, stopping
snake_case_ :int = (
Session().describe_training_job(estimator.latest_training_job.name ).get("""TrainingTimeInSeconds""" , 999_999 )
)
# assert kpis
assert train_runtime <= self.results["train_runtime"]
assert all(t >= self.results["""eval_accuracy"""] for t in eval_accuracy )
assert all(t <= self.results["""eval_loss"""] for t in eval_loss )
# dump tests result into json file to share in PR
with open(f"""{estimator.latest_training_job.name}.json""" , """w""" ) as outfile:
json.dump({"""train_time""": train_runtime, """eval_accuracy""": eval_accuracy, """eval_loss""": eval_loss} , snake_case )
| 66 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tokenizers_available, is_torch_available
__a = {
"configuration_altclip": [
"ALTCLIP_PRETRAINED_CONFIG_ARCHIVE_MAP",
"AltCLIPConfig",
"AltCLIPTextConfig",
"AltCLIPVisionConfig",
],
"processing_altclip": ["AltCLIPProcessor"],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"ALTCLIP_PRETRAINED_MODEL_ARCHIVE_LIST",
"AltCLIPPreTrainedModel",
"AltCLIPModel",
"AltCLIPTextModel",
"AltCLIPVisionModel",
]
if TYPE_CHECKING:
from .configuration_altclip import (
ALTCLIP_PRETRAINED_CONFIG_ARCHIVE_MAP,
AltCLIPConfig,
AltCLIPTextConfig,
AltCLIPVisionConfig,
)
from .processing_altclip import AltCLIPProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_altclip import (
ALTCLIP_PRETRAINED_MODEL_ARCHIVE_LIST,
AltCLIPModel,
AltCLIPPreTrainedModel,
AltCLIPTextModel,
AltCLIPVisionModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 | 1 |
"""simple docstring"""
import argparse
import requests
import torch
# pip3 install salesforce-lavis
# I'm actually installing a slightly modified version: pip3 install git+https://github.com/nielsrogge/LAVIS.git@fix_lavis_float32 (there's also the fix_lavis branch)
# also note: to convert Vicuna checkpoints, we had to include /home/niels/python_projects/checkpoints/FastChat/vicuna-7b in lavis/configs/models/blip2/blip2_instruct_vicuna7b.yaml
# same for Vicuna-13b
from lavis.models import load_model_and_preprocess
from PIL import Image
from transformers import (
AutoTokenizer,
BlipImageProcessor,
InstructBlipConfig,
InstructBlipForConditionalGeneration,
InstructBlipProcessor,
InstructBlipQFormerConfig,
InstructBlipVisionConfig,
LlamaConfig,
LlamaTokenizerFast,
TaConfig,
TaTokenizerFast,
)
from transformers.utils.constants import OPENAI_CLIP_MEAN, OPENAI_CLIP_STD
def A_ ( ):
'''simple docstring'''
snake_case_ :Optional[int] = """https://raw.githubusercontent.com/salesforce/LAVIS/main/docs/_static/Confusing-Pictures.jpg"""
snake_case_ :Optional[Any] = Image.open(requests.get(_lowercase, stream=_lowercase ).raw ).convert("""RGB""" )
return image
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :str = []
# fmt: off
# vision encoder
rename_keys.append(("""visual_encoder.cls_token""", """vision_model.embeddings.class_embedding""") )
rename_keys.append(("""visual_encoder.pos_embed""", """vision_model.embeddings.position_embedding""") )
rename_keys.append(("""visual_encoder.patch_embed.proj.weight""", """vision_model.embeddings.patch_embedding.weight""") )
rename_keys.append(("""visual_encoder.patch_embed.proj.bias""", """vision_model.embeddings.patch_embedding.bias""") )
rename_keys.append(("""ln_vision.weight""", """vision_model.post_layernorm.weight""") )
rename_keys.append(("""ln_vision.bias""", """vision_model.post_layernorm.bias""") )
for i in range(config.vision_config.num_hidden_layers ):
rename_keys.append((f"""visual_encoder.blocks.{i}.norm1.weight""", f"""vision_model.encoder.layers.{i}.layer_norm1.weight""") )
rename_keys.append((f"""visual_encoder.blocks.{i}.norm1.bias""", f"""vision_model.encoder.layers.{i}.layer_norm1.bias""") )
rename_keys.append((f"""visual_encoder.blocks.{i}.norm2.weight""", f"""vision_model.encoder.layers.{i}.layer_norm2.weight""") )
rename_keys.append((f"""visual_encoder.blocks.{i}.norm2.bias""", f"""vision_model.encoder.layers.{i}.layer_norm2.bias""") )
rename_keys.append((f"""visual_encoder.blocks.{i}.attn.qkv.weight""", f"""vision_model.encoder.layers.{i}.self_attn.qkv.weight""") )
rename_keys.append((f"""visual_encoder.blocks.{i}.attn.proj.weight""", f"""vision_model.encoder.layers.{i}.self_attn.projection.weight""",) )
rename_keys.append((f"""visual_encoder.blocks.{i}.attn.proj.bias""", f"""vision_model.encoder.layers.{i}.self_attn.projection.bias""") )
rename_keys.append((f"""visual_encoder.blocks.{i}.mlp.fc1.weight""", f"""vision_model.encoder.layers.{i}.mlp.fc1.weight""") )
rename_keys.append((f"""visual_encoder.blocks.{i}.mlp.fc1.bias""", f"""vision_model.encoder.layers.{i}.mlp.fc1.bias""") )
rename_keys.append((f"""visual_encoder.blocks.{i}.mlp.fc2.weight""", f"""vision_model.encoder.layers.{i}.mlp.fc2.weight""") )
rename_keys.append((f"""visual_encoder.blocks.{i}.mlp.fc2.bias""", f"""vision_model.encoder.layers.{i}.mlp.fc2.bias""") )
# QFormer
rename_keys.append(("""Qformer.bert.embeddings.LayerNorm.weight""", """qformer.embeddings.layernorm.weight""") )
rename_keys.append(("""Qformer.bert.embeddings.LayerNorm.bias""", """qformer.embeddings.layernorm.bias""") )
# fmt: on
return rename_keys
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :str = dct.pop(_lowercase )
snake_case_ :Dict = val
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
for i in range(config.vision_config.num_hidden_layers ):
# read in original q and v biases
snake_case_ :Dict = state_dict.pop(f"""visual_encoder.blocks.{i}.attn.q_bias""" )
snake_case_ :Tuple = state_dict.pop(f"""visual_encoder.blocks.{i}.attn.v_bias""" )
# next, set bias in the state dict
snake_case_ :Union[str, Any] = torch.cat((q_bias, torch.zeros_like(_lowercase, requires_grad=_lowercase ), v_bias) )
snake_case_ :int = qkv_bias
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :List[str] = 364 if """coco""" in model_name else 224
snake_case_ :Tuple = InstructBlipVisionConfig(image_size=_lowercase ).to_dict()
# make sure the models have proper bos_token_id and eos_token_id set (important for generation)
# seems like flan-T5 models don't have bos_token_id properly set?
if "t5-xl" in model_name:
snake_case_ :Optional[Any] = TaConfig.from_pretrained("""google/flan-t5-xl""", dense_act_fn="""gelu""", bos_token_id=1 ).to_dict()
elif "t5-xxl" in model_name:
snake_case_ :List[Any] = TaConfig.from_pretrained("""google/flan-t5-xxl""", dense_act_fn="""gelu""", bos_token_id=1 ).to_dict()
elif "vicuna-7b" in model_name:
snake_case_ :Optional[int] = LlamaConfig.from_pretrained("""decapoda-research/llama-7b-hf""", vocab_size=32001 ).to_dict()
elif "vicuna-13b" in model_name:
snake_case_ :Optional[Any] = LlamaConfig.from_pretrained("""decapoda-research/llama-13b-hf""", vocab_size=32001 ).to_dict()
else:
raise ValueError("""Model name not supported""" )
# the authors add one special "[DEC]" token to the vocab of Q-Former, hence vocab size = 30522 + 1
snake_case_ :Optional[int] = InstructBlipQFormerConfig(vocab_size=30523 ).to_dict()
snake_case_ :List[str] = InstructBlipConfig(vision_config=_lowercase, text_config=_lowercase, qformer_config=_lowercase )
return config, image_size
@torch.no_grad()
def A_ ( _lowercase, _lowercase=None, _lowercase=False ):
'''simple docstring'''
snake_case_ :int = AutoTokenizer.from_pretrained("""bert-base-uncased""", truncation_side="""left""" )
qformer_tokenizer.add_special_tokens({"""bos_token""": """[DEC]"""} )
if "t5" in model_name:
snake_case_ :Optional[int] = TaTokenizerFast.from_pretrained("""google/flan-t5-xl""", truncation_side="""left""" )
elif "vicuna" in model_name:
# the following was used in the original implementation:
# tokenizer = LlamaTokenizer.from_pretrained("huggyllama/llama-7b", use_fast=False, truncation_side="left")
# tokenizer.add_special_tokens({"pad_token": "[PAD]"})
# tokenizer.add_special_tokens({"bos_token": "</s>"})
# tokenizer.add_special_tokens({"eos_token": "</s>"})
# tokenizer.add_special_tokens({"unk_token": "</s>"})
snake_case_ :str = LlamaTokenizerFast.from_pretrained(
"""huggyllama/llama-7b""", truncation_side="""left""", bos_token="""</s>""", unk_token="""</s>""" )
tokenizer.add_special_tokens({"""pad_token""": """[PAD]"""} )
snake_case_, snake_case_ :Union[str, Any] = get_blipa_config(_lowercase )
snake_case_ :Union[str, Any] = InstructBlipForConditionalGeneration(_lowercase ).eval()
snake_case_ :Union[str, Any] = {
"""instructblip-vicuna-7b""": ("""blip2_vicuna_instruct""", """vicuna7b"""),
"""instructblip-vicuna-13b""": ("""blip2_vicuna_instruct""", """vicuna13b"""),
"""instructblip-flan-t5-xl""": ("""blip2_t5_instruct""", """flant5xl"""),
"""instructblip-flan-t5-xxl""": ("""blip2_t5_instruct""", """flant5xxl"""),
}
snake_case_, snake_case_ :List[Any] = model_name_to_original[model_name]
# load original model
print("""Loading original model...""" )
snake_case_ :Optional[Any] = """cuda:1""" if torch.cuda.is_available() else """cpu"""
snake_case_ :Any = """cuda:2""" if torch.cuda.is_available() else """cpu"""
snake_case_, snake_case_, snake_case_ :int = load_model_and_preprocess(
name=_lowercase, model_type=_lowercase, is_eval=_lowercase, device=_lowercase )
original_model.eval()
print("""Done!""" )
# update state dict keys
snake_case_ :int = original_model.state_dict()
snake_case_ :int = create_rename_keys(_lowercase )
for src, dest in rename_keys:
rename_key(_lowercase, _lowercase, _lowercase )
# some keys can be renamed efficiently
for key, val in state_dict.copy().items():
snake_case_ :Optional[Any] = state_dict.pop(_lowercase )
if key.startswith("""Qformer.bert""" ):
snake_case_ :str = key.replace("""Qformer.bert""", """qformer""" )
if "attention.self" in key:
snake_case_ :int = key.replace("""self""", """attention""" )
if "llm_proj" in key:
snake_case_ :Dict = key.replace("""llm_proj""", """language_projection""" )
if "t5_proj" in key:
snake_case_ :Dict = key.replace("""t5_proj""", """language_projection""" )
if key.startswith("""llm_model""" ):
snake_case_ :Union[str, Any] = key.replace("""llm_model""", """language_model""" )
if key.startswith("""t5""" ):
snake_case_ :List[str] = key.replace("""t5""", """language""" )
snake_case_ :List[str] = val
# read in qv biases
read_in_q_v_bias(_lowercase, _lowercase )
# note: weights get loaded in torch.float32 by default
hf_model.load_state_dict(_lowercase, strict=_lowercase )
snake_case_ :Optional[Any] = load_demo_image()
snake_case_ :Tuple = """What is unusual about this image?"""
# create processor
snake_case_ :Tuple = BlipImageProcessor(
size={"""height""": image_size, """width""": image_size}, image_mean=_lowercase, image_std=_lowercase )
snake_case_ :Tuple = InstructBlipProcessor(
image_processor=_lowercase, tokenizer=_lowercase, qformer_tokenizer=_lowercase, )
snake_case_ :Tuple = processor(images=_lowercase, text=_lowercase, return_tensors="""pt""" ).to(_lowercase )
# make sure processor creates exact same pixel values
snake_case_ :Optional[Any] = vis_processors["""eval"""](_lowercase ).unsqueeze(0 ).to(_lowercase )
snake_case_ :List[str] = inputs.pixel_values
assert torch.allclose(original_pixel_values.to(pixel_values.device ), _lowercase )
original_model.to(_lowercase )
hf_model.to(_lowercase )
with torch.no_grad():
if "vicuna" in model_name:
snake_case_ :List[Any] = original_model({"""image""": original_pixel_values, """text_input""": [prompt]} ).logits
snake_case_ :Union[str, Any] = hf_model(**_lowercase ).logits
else:
snake_case_ :Dict = original_model(
{"""image""": original_pixel_values, """text_input""": [prompt], """text_output""": ["""\n"""]} ).logits
snake_case_ :Dict = tokenizer("""\n""", return_tensors="""pt""" ).input_ids.to(_lowercase )
snake_case_ :List[str] = label_input_ids.masked_fill(label_input_ids == tokenizer.pad_token_id, -100 )
snake_case_ :List[Any] = hf_model(**_lowercase, labels=_lowercase ).logits
print("""First values of original logits:""", original_logits[0, :3, :3] )
print("""First values of HF logits:""", logits[0, :3, :3] )
# assert values
assert original_logits.shape == logits.shape
snake_case_ :Any = 1e-4 if """vicuna""" in model_name else 1e-5
assert torch.allclose(original_logits.to(logits.device ), _lowercase, atol=_lowercase )
print("""Looks ok!""" )
print("""Generating with original model...""" )
snake_case_ :Dict = original_model.generate({"""image""": original_pixel_values, """prompt""": prompt}, num_beams=5 )
# important: we need to cast the weights of the HF model to the appropriate type
print("""Generating with HF model...""" )
snake_case_ :str = hf_model.generate(
**_lowercase, do_sample=_lowercase, num_beams=5, max_length=256, min_length=1, top_p=0.9, repetition_penalty=1.5, length_penalty=1.0, temperature=1, )
if "vicuna" in model_name:
# convert output id 0 to 2 (eos_token_id)
# TODO add this in the generate method?
snake_case_ :Optional[int] = 2
print("""Original generation:""", _lowercase )
snake_case_ :str = processor.batch_decode(_lowercase, skip_special_tokens=_lowercase )
snake_case_ :Any = [text.strip() for text in output_text]
print("""HF generation:""", _lowercase )
if pytorch_dump_folder_path is not None:
processor.save_pretrained(_lowercase )
hf_model.save_pretrained(_lowercase )
if push_to_hub:
processor.push_to_hub(f"""Salesforce/{model_name}""" )
hf_model.push_to_hub(f"""Salesforce/{model_name}""" )
if __name__ == "__main__":
__a = argparse.ArgumentParser()
__a = [
"instructblip-vicuna-7b",
"instructblip-vicuna-13b",
"instructblip-flan-t5-xl",
"instructblip-flan-t5-xxl",
]
parser.add_argument(
"--model_name",
default="instructblip-flan-t5-xl",
choices=choices,
type=str,
help="Path to hf config.json of model to convert",
)
parser.add_argument("--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model.")
parser.add_argument(
"--push_to_hub",
action="store_true",
help="Whether to push the model and processor to the hub after converting",
)
__a = parser.parse_args()
convert_blipa_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 66 |
"""simple docstring"""
import argparse
import gdown
import numpy as np
import torch
from huggingface_hub import hf_hub_download
from transformers import (
CLIPTokenizer,
CLIPTokenizerFast,
VideoMAEImageProcessor,
XCLIPConfig,
XCLIPModel,
XCLIPProcessor,
XCLIPTextConfig,
XCLIPVisionConfig,
)
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :int = XCLIPTextConfig()
# derive patch size from model name
snake_case_ :Union[str, Any] = model_name.find("""patch""" )
snake_case_ :List[str] = int(model_name[start_idx + len("""patch""" ) : start_idx + len("""patch""" ) + 2] )
snake_case_ :Any = XCLIPVisionConfig(patch_size=_lowercase, num_frames=_lowercase )
if "large" in model_name:
snake_case_ :Optional[Any] = 768
snake_case_ :Union[str, Any] = 3072
snake_case_ :Any = 12
snake_case_ :Any = 1024
snake_case_ :str = 4096
snake_case_ :Union[str, Any] = 16
snake_case_ :Union[str, Any] = 24
snake_case_ :Tuple = 768
snake_case_ :Any = 3072
if model_name == "xclip-large-patch14-16-frames":
snake_case_ :Any = 336
snake_case_ :Any = XCLIPConfig.from_text_vision_configs(_lowercase, _lowercase )
if "large" in model_name:
snake_case_ :List[Any] = 768
return config
def A_ ( _lowercase ):
'''simple docstring'''
if name == "token_embedding.weight":
snake_case_ :Optional[Any] = name.replace("""token_embedding.weight""", """text_model.embeddings.token_embedding.weight""" )
if name == "positional_embedding":
snake_case_ :Tuple = name.replace("""positional_embedding""", """text_model.embeddings.position_embedding.weight""" )
if "ln_1" in name:
snake_case_ :Dict = name.replace("""ln_1""", """layer_norm1""" )
if "ln_2" in name:
snake_case_ :str = name.replace("""ln_2""", """layer_norm2""" )
if "c_fc" in name:
snake_case_ :str = name.replace("""c_fc""", """fc1""" )
if "c_proj" in name:
snake_case_ :int = name.replace("""c_proj""", """fc2""" )
if name.startswith("""transformer.resblocks""" ):
snake_case_ :Union[str, Any] = name.replace("""transformer.resblocks""", """text_model.encoder.layers""" )
if "attn.out_proj" in name and "message" not in name:
snake_case_ :Union[str, Any] = name.replace("""attn.out_proj""", """self_attn.out_proj""" )
if "ln_final" in name:
snake_case_ :Union[str, Any] = name.replace("""ln_final""", """text_model.final_layer_norm""" )
# visual encoder
if name == "visual.class_embedding":
snake_case_ :Any = name.replace("""visual.class_embedding""", """vision_model.embeddings.class_embedding""" )
if name == "visual.positional_embedding":
snake_case_ :Optional[int] = name.replace("""visual.positional_embedding""", """vision_model.embeddings.position_embedding.weight""" )
if name.startswith("""visual.transformer.resblocks""" ):
snake_case_ :Union[str, Any] = name.replace("""visual.transformer.resblocks""", """vision_model.encoder.layers""" )
if "visual.conv1" in name:
snake_case_ :int = name.replace("""visual.conv1""", """vision_model.embeddings.patch_embedding""" )
if "visual.ln_pre" in name:
snake_case_ :Any = name.replace("""visual.ln_pre""", """vision_model.pre_layernorm""" )
if "visual.ln_post" in name:
snake_case_ :str = name.replace("""visual.ln_post""", """vision_model.post_layernorm""" )
if "visual.proj" in name:
snake_case_ :Union[str, Any] = name.replace("""visual.proj""", """visual_projection.weight""" )
if "text_projection" in name:
snake_case_ :Dict = name.replace("""text_projection""", """text_projection.weight""" )
# things on top
if "prompts_visual_proj" in name:
snake_case_ :List[str] = name.replace("""prompts_visual_proj""", """prompts_visual_projection""" )
if "prompts_visual_ln" in name:
snake_case_ :Dict = name.replace("""prompts_visual_ln""", """prompts_visual_layernorm""" )
# mit
if name == "mit.positional_embedding":
snake_case_ :str = name.replace("""positional""", """position""" )
if name.startswith("""mit.resblocks""" ):
snake_case_ :Dict = name.replace("""mit.resblocks""", """mit.encoder.layers""" )
# prompts generator
if name.startswith("""prompts_generator.norm""" ):
snake_case_ :Union[str, Any] = name.replace("""prompts_generator.norm""", """prompts_generator.layernorm""" )
return name
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
for key in orig_state_dict.copy().keys():
snake_case_ :Dict = orig_state_dict.pop(_lowercase )
if "attn.in_proj" in key:
snake_case_ :Optional[Any] = key.split(""".""" )
if key.startswith("""visual""" ):
snake_case_ :Any = key_split[3]
snake_case_ :Optional[Any] = config.vision_config.hidden_size
if "message_attn" in key:
if "weight" in key:
snake_case_ :str = val[
:dim, :
]
snake_case_ :Optional[int] = val[
dim : dim * 2, :
]
snake_case_ :Union[str, Any] = val[
-dim:, :
]
else:
snake_case_ :Dict = val[
:dim
]
snake_case_ :Optional[int] = val[
dim : dim * 2
]
snake_case_ :Optional[int] = val[
-dim:
]
else:
if "weight" in key:
snake_case_ :Optional[Any] = val[
:dim, :
]
snake_case_ :List[str] = val[
dim : dim * 2, :
]
snake_case_ :Dict = val[
-dim:, :
]
else:
snake_case_ :Union[str, Any] = val[:dim]
snake_case_ :Union[str, Any] = val[
dim : dim * 2
]
snake_case_ :Union[str, Any] = val[-dim:]
elif key.startswith("""mit""" ):
snake_case_ :Tuple = key_split[2]
snake_case_ :Union[str, Any] = config.vision_config.mit_hidden_size
if "weight" in key:
snake_case_ :Optional[int] = val[:dim, :]
snake_case_ :Optional[int] = val[dim : dim * 2, :]
snake_case_ :str = val[-dim:, :]
else:
snake_case_ :str = val[:dim]
snake_case_ :Any = val[dim : dim * 2]
snake_case_ :int = val[-dim:]
else:
snake_case_ :Tuple = key_split[2]
snake_case_ :Any = config.text_config.hidden_size
if "weight" in key:
snake_case_ :Dict = val[:dim, :]
snake_case_ :Dict = val[
dim : dim * 2, :
]
snake_case_ :List[str] = val[-dim:, :]
else:
snake_case_ :Any = val[:dim]
snake_case_ :Tuple = val[
dim : dim * 2
]
snake_case_ :List[str] = val[-dim:]
else:
snake_case_ :Optional[int] = rename_key(_lowercase )
if new_key_name in ["visual_projection.weight", "text_projection.weight"]:
snake_case_ :Optional[Any] = val.T
snake_case_ :Tuple = val
return orig_state_dict
def A_ ( _lowercase ):
'''simple docstring'''
if num_frames == 8:
snake_case_ :str = """eating_spaghetti_8_frames.npy"""
elif num_frames == 16:
snake_case_ :int = """eating_spaghetti.npy"""
elif num_frames == 32:
snake_case_ :List[str] = """eating_spaghetti_32_frames.npy"""
snake_case_ :int = hf_hub_download(
repo_id="""hf-internal-testing/spaghetti-video""", filename=_lowercase, repo_type="""dataset""", )
snake_case_ :Union[str, Any] = np.load(_lowercase )
return list(_lowercase )
def A_ ( _lowercase, _lowercase=None, _lowercase=False ):
'''simple docstring'''
snake_case_ :List[Any] = {
# fully supervised kinetics-400 checkpoints
"""xclip-base-patch32""": """https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_32_8.pth""",
"""xclip-base-patch32-16-frames""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_32_16.pth"""
),
"""xclip-base-patch16""": """https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_16_8.pth""",
"""xclip-base-patch16-16-frames""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_16_16.pth"""
),
"""xclip-large-patch14""": """https://drive.google.com/u/0/uc?id=1NUOImq0o5DlQTST17iIP3vG7DgmHQuCx&export=download&confirm=t&uuid=b26caedc-88e2-473e-830a-9d158b653cdb""",
"""xclip-large-patch14-16-frames""": """https://drive.google.com/u/0/uc?id=1FOYgnJc097OJ4lGwtRCCydQyVPJEOH7d&export=download&confirm=t&uuid=538fa810-e671-4050-b385-9a623f89804f""",
# fully supervised kinetics-600 checkpoints
"""xclip-base-patch16-kinetics-600""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k600_16_8.pth"""
),
"""xclip-base-patch16-kinetics-600-16-frames""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k600_16_16.pth"""
),
"""xclip-large-patch14-kinetics-600""": """https://drive.google.com/u/0/uc?id=1FV8C1INuM91sLAN4ImjzePLIlpMSihwV&export=download&confirm=t&uuid=141d4977-4a65-44ae-864f-4b0c19f838be""",
# few shot
"""xclip-base-patch16-hmdb-2-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_2.pth"""
),
"""xclip-base-patch16-hmdb-4-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_4.pth"""
),
"""xclip-base-patch16-hmdb-8-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_8.pth"""
),
"""xclip-base-patch16-hmdb-16-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_16.pth"""
),
"""xclip-base-patch16-ucf-2-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_2.pth"""
),
"""xclip-base-patch16-ucf-4-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_4.pth"""
),
"""xclip-base-patch16-ucf-8-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_8.pth"""
),
"""xclip-base-patch16-ucf-16-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_16.pth"""
),
# zero shot
"""xclip-base-patch16-zero-shot""": """https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/zero.pth""",
}
snake_case_ :Optional[int] = model_to_url[model_name]
snake_case_ :int = 8
if "16-frames" in model_name:
snake_case_ :List[Any] = 16
elif "shot" in model_name:
snake_case_ :Dict = 32
snake_case_ :Optional[int] = get_xclip_config(_lowercase, _lowercase )
snake_case_ :Optional[Any] = XCLIPModel(_lowercase )
model.eval()
if "drive" in checkpoint_url:
snake_case_ :List[str] = """pytorch_model.bin"""
gdown.cached_download(_lowercase, _lowercase, quiet=_lowercase )
snake_case_ :List[Any] = torch.load(_lowercase, map_location="""cpu""" )["""model"""]
else:
snake_case_ :Tuple = torch.hub.load_state_dict_from_url(_lowercase )["""model"""]
snake_case_ :Union[str, Any] = convert_state_dict(_lowercase, _lowercase )
snake_case_ :str = XCLIPModel(_lowercase )
snake_case_, snake_case_ :Optional[int] = model.load_state_dict(_lowercase, strict=_lowercase )
assert missing_keys == ["text_model.embeddings.position_ids", "vision_model.embeddings.position_ids"]
model.eval()
snake_case_ :List[str] = 336 if model_name == """xclip-large-patch14-16-frames""" else 224
snake_case_ :List[Any] = VideoMAEImageProcessor(size=_lowercase )
snake_case_ :Any = CLIPTokenizer.from_pretrained("""openai/clip-vit-base-patch32""" )
snake_case_ :str = CLIPTokenizerFast.from_pretrained("""openai/clip-vit-base-patch32""" )
snake_case_ :Optional[Any] = XCLIPProcessor(image_processor=_lowercase, tokenizer=_lowercase )
snake_case_ :Optional[int] = prepare_video(_lowercase )
snake_case_ :Optional[Any] = processor(
text=["""playing sports""", """eating spaghetti""", """go shopping"""], videos=_lowercase, return_tensors="""pt""", padding=_lowercase )
print("""Shape of pixel values:""", inputs.pixel_values.shape )
with torch.no_grad():
snake_case_ :List[Any] = model(**_lowercase )
# Verify outputs
snake_case_ :List[Any] = outputs.logits_per_video
snake_case_ :Any = logits_per_video.softmax(dim=1 )
print("""Probs:""", _lowercase )
# kinetics-400
if model_name == "xclip-base-patch32":
snake_case_ :Union[str, Any] = torch.tensor([[0.0019, 0.9951, 0.0030]] )
elif model_name == "xclip-base-patch32-16-frames":
snake_case_ :str = torch.tensor([[7.09_99e-04, 9.98_83e-01, 4.55_80e-04]] )
elif model_name == "xclip-base-patch16":
snake_case_ :Tuple = torch.tensor([[0.0083, 0.9681, 0.0236]] )
elif model_name == "xclip-base-patch16-16-frames":
snake_case_ :Any = torch.tensor([[7.69_37e-04, 9.97_28e-01, 1.94_73e-03]] )
elif model_name == "xclip-large-patch14":
snake_case_ :str = torch.tensor([[0.0062, 0.9864, 0.0075]] )
elif model_name == "xclip-large-patch14-16-frames":
snake_case_ :Tuple = torch.tensor([[3.38_77e-04, 9.99_37e-01, 2.88_88e-04]] )
# kinetics-600
elif model_name == "xclip-base-patch16-kinetics-600":
snake_case_ :List[Any] = torch.tensor([[0.0555, 0.8914, 0.0531]] )
elif model_name == "xclip-base-patch16-kinetics-600-16-frames":
snake_case_ :Union[str, Any] = torch.tensor([[3.85_54e-04, 9.99_29e-01, 3.27_54e-04]] )
elif model_name == "xclip-large-patch14-kinetics-600":
snake_case_ :List[Any] = torch.tensor([[0.0036, 0.9920, 0.0045]] )
# few shot
elif model_name == "xclip-base-patch16-hmdb-2-shot":
snake_case_ :Dict = torch.tensor([[7.18_90e-06, 9.99_94e-01, 5.65_59e-05]] )
elif model_name == "xclip-base-patch16-hmdb-4-shot":
snake_case_ :Union[str, Any] = torch.tensor([[1.03_20e-05, 9.99_93e-01, 6.24_35e-05]] )
elif model_name == "xclip-base-patch16-hmdb-8-shot":
snake_case_ :str = torch.tensor([[4.13_77e-06, 9.99_90e-01, 9.83_86e-05]] )
elif model_name == "xclip-base-patch16-hmdb-16-shot":
snake_case_ :str = torch.tensor([[4.13_47e-05, 9.99_62e-01, 3.34_11e-04]] )
elif model_name == "xclip-base-patch16-ucf-2-shot":
snake_case_ :int = torch.tensor([[8.58_57e-05, 9.99_28e-01, 6.32_91e-04]] )
elif model_name == "xclip-base-patch16-ucf-4-shot":
snake_case_ :Optional[int] = torch.tensor([[8.58_57e-05, 9.99_28e-01, 6.32_91e-04]] )
elif model_name == "xclip-base-patch16-ucf-8-shot":
snake_case_ :Any = torch.tensor([[0.0027, 0.9904, 0.0070]] )
elif model_name == "xclip-base-patch16-ucf-16-shot":
snake_case_ :Tuple = torch.tensor([[9.82_19e-04, 9.95_93e-01, 3.08_63e-03]] )
# zero shot
elif model_name == "xclip-base-patch16-zero-shot":
snake_case_ :Union[str, Any] = torch.tensor([[3.50_82e-04, 9.97_85e-01, 1.79_66e-03]] )
else:
raise ValueError(f"""Model name {model_name} not supported""" )
assert torch.allclose(_lowercase, _lowercase, atol=1e-3 )
print("""Looks ok!""" )
if pytorch_dump_folder_path is not None:
print(f"""Saving model {model_name} to {pytorch_dump_folder_path}""" )
model.save_pretrained(_lowercase )
if push_to_hub:
print("""Pushing model, processor and slow tokenizer files to the hub...""" )
model.push_to_hub(_lowercase, organization="""nielsr""" )
processor.push_to_hub(_lowercase, organization="""nielsr""" )
slow_tokenizer.push_to_hub(_lowercase, organization="""nielsr""" )
if __name__ == "__main__":
__a = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--model_name",
default="xclip-base-patch32",
type=str,
help="Name of the model.",
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model directory."
)
parser.add_argument(
"--push_to_hub", action="store_true", help="Whether or not to push the converted model to the 🤗 hub."
)
__a = parser.parse_args()
convert_xclip_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 66 | 1 |
"""simple docstring"""
import argparse
import re
from flax.traverse_util import flatten_dict, unflatten_dict
from tax import checkpoints
from transformers import SwitchTransformersConfig, SwitchTransformersForConditionalGeneration
from transformers.modeling_flax_pytorch_utils import load_flax_weights_in_pytorch_model
from transformers.utils import logging
logging.set_verbosity_info()
# should not include what is already done by the `from_pt` argument
__a = {
"/attention/": "/0/SelfAttention/",
"/self_attention/": "/0/SelfAttention/",
"/encoder_decoder_attention/": "/1/EncDecAttention/",
"value": "v",
"query": "q",
"key": "k",
"out": "o",
"pre_self_attention_layer_norm": "0/layer_norm",
"pre_cross_attention_layer_norm": "1/layer_norm",
"pre_attention_layer_norm": "0/layer_norm", # previously 1, but seems wrong
"token_embedder": "shared",
"encoder_norm": "final_layer_norm",
"decoder_norm": "final_layer_norm",
"relpos_bias/rel_embedding": "block/0/layer/0/SelfAttention/relative_attention_bias/weight",
"router/router_weights/w/": "router/classifier/",
"roer/roer_weights/w/": "router/classifier/",
"logits_dense": "lm_head",
}
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Optional[Any] = list(s_dict.keys() )
for key in keys:
snake_case_ :Dict = r""".*/layers_(\d+)"""
snake_case_ :Dict = key
if re.match(_lowercase, _lowercase ):
snake_case_ :Dict = re.sub(r"""layers_(\d+)""", r"""block/\1/layer""", _lowercase )
snake_case_ :List[Any] = r"""(encoder|decoder)\/"""
if re.match(_lowercase, _lowercase ):
snake_case_ :Any = re.match(_lowercase, _lowercase ).groups()
if groups[0] == "encoder":
snake_case_ :str = re.sub(r"""/mlp/""", r"""/1/mlp/""", _lowercase )
snake_case_ :Optional[Any] = re.sub(r"""/pre_mlp_layer_norm/""", r"""/1/layer_norm/""", _lowercase )
elif groups[0] == "decoder":
snake_case_ :int = re.sub(r"""/mlp/""", r"""/2/mlp/""", _lowercase )
snake_case_ :Optional[Any] = re.sub(r"""/pre_mlp_layer_norm/""", r"""/2/layer_norm/""", _lowercase )
# 2. Convert other classic mappings
for old_key, temp_key in MOE_LAYER_NAME_MAPPING.items():
if old_key in new_key:
snake_case_ :Union[str, Any] = new_key.replace(_lowercase, _lowercase )
print(f"""{key} -> {new_key}""" )
snake_case_ :List[Any] = s_dict.pop(_lowercase )
if "encoder/block/0/layer/0/SelfAttention/relative_attention_bias/weight" in s_dict:
snake_case_ :List[Any] = s_dict[
"""encoder/block/0/layer/0/SelfAttention/relative_attention_bias/weight"""
].T
if "decoder/block/0/layer/0/SelfAttention/relative_attention_bias/weight" in s_dict:
snake_case_ :Tuple = s_dict[
"""decoder/block/0/layer/0/SelfAttention/relative_attention_bias/weight"""
].T
# 3. Take extra care of the EXPERTS layer
for key in list(s_dict.keys() ):
if "expert" in key:
snake_case_ :Tuple = s_dict[key].shape[0]
snake_case_ :int = s_dict[key]
for idx in range(_lowercase ):
snake_case_ :Union[str, Any] = expert_weihts[idx]
print(f"""{key} -> {key.replace('expert/', 'nested fstring' )}""" )
s_dict.pop(_lowercase )
return s_dict
__a = {
"NUM_ENCODER_LAYERS": "num_layers",
"NUM_DECODER_LAYERS": "num_decoder_layers",
"NUM_HEADS": "num_heads",
"HEAD_DIM": "d_kv",
"EMBED_DIM": "d_model",
"MLP_DIM": "d_ff",
"NUM_SELECTED_EXPERTS": "num_selected_experts",
"NUM_ENCODER_SPARSE_LAYERS": "num_sparse_encoder_layers",
"NUM_DECODER_SPARSE_LAYERS": "num_sparse_decoder_layers",
"dense.MlpBlock.activations": "feed_forward_proj",
}
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
import regex as re
with open(_lowercase, """r""" ) as f:
snake_case_ :Tuple = f.read()
snake_case_ :Dict = re.findall(r"""(.*) = ([0-9.]*)""", _lowercase )
snake_case_ :Union[str, Any] = {}
for param, value in regex_match:
if param in GIN_TO_CONFIG_MAPPING and value != "":
snake_case_ :int = float(_lowercase ) if """.""" in value else int(_lowercase )
snake_case_ :Dict = re.findall(r"""(.*activations) = \(\'(.*)\',\)""", _lowercase )[0]
snake_case_ :Optional[int] = str(activation[1] )
snake_case_ :Dict = num_experts
snake_case_ :int = SwitchTransformersConfig(**_lowercase )
return config
def A_ ( _lowercase, _lowercase, _lowercase=None, _lowercase="./", _lowercase=8 ):
'''simple docstring'''
print(f"""Loading flax weights from : {flax_checkpoint_path}""" )
snake_case_ :List[Any] = checkpoints.load_tax_checkpoint(_lowercase )
if gin_file is not None:
snake_case_ :Dict = convert_gin_to_config(_lowercase, _lowercase )
else:
snake_case_ :Optional[Any] = SwitchTransformersConfig.from_pretrained(_lowercase )
snake_case_ :Union[str, Any] = SwitchTransformersForConditionalGeneration(_lowercase )
snake_case_ :Tuple = flax_params["""target"""]
snake_case_ :Any = flatten_dict(_lowercase, sep="""/""" )
snake_case_ :int = rename_keys(_lowercase )
snake_case_ :List[Any] = unflatten_dict(_lowercase, sep="""/""" )
# Load the flax params in the PT model
load_flax_weights_in_pytorch_model(_lowercase, _lowercase )
print(f"""Save PyTorch model to {pytorch_dump_path}""" )
pt_model.save_pretrained(_lowercase )
if __name__ == "__main__":
__a = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--switch_t5x_checkpoint_path",
default=None,
type=str,
required=True,
help=(
"The config json file corresponding to the pre-trained SwitchTransformers model. \nThis specifies the"
" model architecture. If not provided, a `gin_file` has to be provided."
),
)
parser.add_argument(
"--gin_file",
default=None,
type=str,
required=False,
help="Path to the gin config file. If not provided, a `config_file` has to be passed ",
)
parser.add_argument(
"--config_name", default=None, type=str, required=False, help="Config name of SwitchTransformers model."
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, required=True, help="Path to the output pytorch model."
)
parser.add_argument("--num_experts", default=8, type=int, required=False, help="Number of experts")
__a = parser.parse_args()
convert_flax_checkpoint_to_pytorch(
args.switch_tax_checkpoint_path,
args.config_name,
args.gin_file,
args.pytorch_dump_folder_path,
args.num_experts,
)
| 66 |
"""simple docstring"""
import unittest
import numpy as np
from transformers import BertConfig, is_flax_available
from transformers.testing_utils import require_flax, slow
from ...test_modeling_flax_common import FlaxModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask
if is_flax_available():
from transformers.models.bert.modeling_flax_bert import (
FlaxBertForMaskedLM,
FlaxBertForMultipleChoice,
FlaxBertForNextSentencePrediction,
FlaxBertForPreTraining,
FlaxBertForQuestionAnswering,
FlaxBertForSequenceClassification,
FlaxBertForTokenClassification,
FlaxBertModel,
)
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def __init__( self: List[Any] , snake_case: List[str] , snake_case: Optional[Any]=13 , snake_case: List[str]=7 , snake_case: Dict=True , snake_case: List[str]=True , snake_case: Optional[int]=True , snake_case: Any=True , snake_case: Optional[Any]=99 , snake_case: Tuple=32 , snake_case: Tuple=5 , snake_case: Dict=4 , snake_case: Optional[Any]=37 , snake_case: Union[str, Any]="gelu" , snake_case: Tuple=0.1 , snake_case: List[Any]=0.1 , snake_case: List[str]=512 , snake_case: Optional[int]=16 , snake_case: int=2 , snake_case: List[Any]=0.0_2 , snake_case: Union[str, Any]=4 , ) -> List[str]:
snake_case_ :Dict = parent
snake_case_ :Any = batch_size
snake_case_ :Any = seq_length
snake_case_ :List[str] = is_training
snake_case_ :Optional[Any] = use_attention_mask
snake_case_ :Dict = use_token_type_ids
snake_case_ :Union[str, Any] = use_labels
snake_case_ :str = vocab_size
snake_case_ :int = hidden_size
snake_case_ :List[str] = num_hidden_layers
snake_case_ :Dict = num_attention_heads
snake_case_ :Any = intermediate_size
snake_case_ :Tuple = hidden_act
snake_case_ :int = hidden_dropout_prob
snake_case_ :Optional[Any] = attention_probs_dropout_prob
snake_case_ :Any = max_position_embeddings
snake_case_ :Union[str, Any] = type_vocab_size
snake_case_ :Optional[int] = type_sequence_label_size
snake_case_ :Union[str, Any] = initializer_range
snake_case_ :Tuple = num_choices
def lowerCAmelCase_ ( self: Tuple ) -> str:
snake_case_ :Optional[Any] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case_ :Union[str, Any] = None
if self.use_attention_mask:
snake_case_ :str = random_attention_mask([self.batch_size, self.seq_length] )
snake_case_ :Any = None
if self.use_token_type_ids:
snake_case_ :List[str] = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
snake_case_ :int = BertConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , is_decoder=snake_case , initializer_range=self.initializer_range , )
return config, input_ids, token_type_ids, attention_mask
def lowerCAmelCase_ ( self: Optional[int] ) -> int:
snake_case_ :str = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_, snake_case_ :Optional[int] = config_and_inputs
snake_case_ :Union[str, Any] = {"""input_ids""": input_ids, """token_type_ids""": token_type_ids, """attention_mask""": attention_mask}
return config, inputs_dict
def lowerCAmelCase_ ( self: Optional[Any] ) -> Any:
snake_case_ :int = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_, snake_case_ :Dict = config_and_inputs
snake_case_ :Union[str, Any] = True
snake_case_ :Optional[int] = floats_tensor([self.batch_size, self.seq_length, self.hidden_size] )
snake_case_ :Tuple = ids_tensor([self.batch_size, self.seq_length] , vocab_size=2 )
return (
config,
input_ids,
attention_mask,
encoder_hidden_states,
encoder_attention_mask,
)
@require_flax
class lowerCamelCase ( _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[str] = True
_A : Dict = (
(
FlaxBertModel,
FlaxBertForPreTraining,
FlaxBertForMaskedLM,
FlaxBertForMultipleChoice,
FlaxBertForQuestionAnswering,
FlaxBertForNextSentencePrediction,
FlaxBertForSequenceClassification,
FlaxBertForTokenClassification,
FlaxBertForQuestionAnswering,
)
if is_flax_available()
else ()
)
def lowerCAmelCase_ ( self: int ) -> List[str]:
snake_case_ :Any = FlaxBertModelTester(self )
@slow
def lowerCAmelCase_ ( self: List[str] ) -> Dict:
# Only check this for base model, not necessary for all model classes.
# This will also help speed-up tests.
snake_case_ :Dict = FlaxBertModel.from_pretrained("""bert-base-cased""" )
snake_case_ :Dict = model(np.ones((1, 1) ) )
self.assertIsNotNone(snake_case )
| 66 | 1 |
"""simple docstring"""
from __future__ import annotations
from typing import Any
def A_ ( _lowercase ):
'''simple docstring'''
create_state_space_tree(_lowercase, [], 0 )
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
if index == len(_lowercase ):
print(_lowercase )
return
create_state_space_tree(_lowercase, _lowercase, index + 1 )
current_subsequence.append(sequence[index] )
create_state_space_tree(_lowercase, _lowercase, index + 1 )
current_subsequence.pop()
if __name__ == "__main__":
__a = [3, 1, 2, 4]
generate_all_subsequences(seq)
seq.clear()
seq.extend(["A", "B", "C"])
generate_all_subsequences(seq)
| 66 |
"""simple docstring"""
import math
class lowerCamelCase :
'''simple docstring'''
def lowerCAmelCase_ ( self: Tuple , snake_case: list[list[float]] , snake_case: list[int] ) -> int:
snake_case_ :Any = 0.0
snake_case_ :Tuple = 0.0
for i in range(len(snake_case ) ):
da += math.pow((sample[i] - weights[0][i]) , 2 )
da += math.pow((sample[i] - weights[1][i]) , 2 )
return 0 if da > da else 1
return 0
def lowerCAmelCase_ ( self: Optional[int] , snake_case: list[list[int | float]] , snake_case: list[int] , snake_case: int , snake_case: float ) -> list[list[int | float]]:
for i in range(len(snake_case ) ):
weights[j][i] += alpha * (sample[i] - weights[j][i])
return weights
def A_ ( ):
'''simple docstring'''
snake_case_ :Dict = [[1, 1, 0, 0], [0, 0, 0, 1], [1, 0, 0, 0], [0, 0, 1, 1]]
# weight initialization ( n, C )
snake_case_ :List[Any] = [[0.2, 0.6, 0.5, 0.9], [0.8, 0.4, 0.7, 0.3]]
# training
snake_case_ :Optional[Any] = SelfOrganizingMap()
snake_case_ :Dict = 3
snake_case_ :Dict = 0.5
for _ in range(_lowercase ):
for j in range(len(_lowercase ) ):
# training sample
snake_case_ :List[Any] = training_samples[j]
# Compute the winning vector
snake_case_ :Optional[int] = self_organizing_map.get_winner(_lowercase, _lowercase )
# Update the winning vector
snake_case_ :List[str] = self_organizing_map.update(_lowercase, _lowercase, _lowercase, _lowercase )
# classify test sample
snake_case_ :str = [0, 0, 0, 1]
snake_case_ :List[Any] = self_organizing_map.get_winner(_lowercase, _lowercase )
# results
print(f"""Clusters that the test sample belongs to : {winner}""" )
print(f"""Weights that have been trained : {weights}""" )
# running the main() function
if __name__ == "__main__":
main()
| 66 | 1 |
"""simple docstring"""
from typing import Dict, List, Optional, Union
import numpy as np
from .feature_extraction_utils import BatchFeature, FeatureExtractionMixin
from .utils import PaddingStrategy, TensorType, is_tf_tensor, is_torch_tensor, logging, to_numpy
__a = logging.get_logger(__name__)
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
def __init__( self: List[Any] , snake_case: int , snake_case: int , snake_case: float , **snake_case: Optional[int] ) -> Optional[Any]:
snake_case_ :List[Any] = feature_size
snake_case_ :Tuple = sampling_rate
snake_case_ :Optional[int] = padding_value
snake_case_ :Dict = kwargs.pop("""padding_side""" , """right""" )
snake_case_ :List[Any] = kwargs.pop("""return_attention_mask""" , snake_case )
super().__init__(**snake_case )
def lowerCAmelCase_ ( self: List[str] , snake_case: Union[
BatchFeature,
List[BatchFeature],
Dict[str, BatchFeature],
Dict[str, List[BatchFeature]],
List[Dict[str, BatchFeature]],
] , snake_case: Union[bool, str, PaddingStrategy] = True , snake_case: Optional[int] = None , snake_case: bool = False , snake_case: Optional[int] = None , snake_case: Optional[bool] = None , snake_case: Optional[Union[str, TensorType]] = None , ) -> BatchFeature:
# If we have a list of dicts, let's convert it in a dict of lists
# We do this to allow using this method as a collate_fn function in PyTorch Dataloader
if isinstance(snake_case , (list, tuple) ) and isinstance(processed_features[0] , (dict, BatchFeature) ):
snake_case_ :int = {
key: [example[key] for example in processed_features] for key in processed_features[0].keys()
}
# The model's main input name, usually `input_values`, has be passed for padding
if self.model_input_names[0] not in processed_features:
raise ValueError(
"""You should supply an instance of `transformers.BatchFeature` or list of `transformers.BatchFeature`"""
f""" to this method that includes {self.model_input_names[0]}, but you provided"""
f""" {list(processed_features.keys() )}""" )
snake_case_ :Union[str, Any] = processed_features[self.model_input_names[0]]
snake_case_ :Optional[int] = (
return_attention_mask if return_attention_mask is not None else self.return_attention_mask
)
if len(snake_case ) == 0:
if return_attention_mask:
snake_case_ :Optional[int] = []
return processed_features
# If we have PyTorch/TF tensors or lists as inputs, we cast them as Numpy arrays
# and rebuild them afterwards if no return_tensors is specified
# Note that we lose the specific device the tensor may be on for PyTorch
snake_case_ :Optional[Any] = required_input[0]
if isinstance(snake_case , (list, tuple) ):
# first_element might be an empty list/tuple in some edge cases so we grab the first non empty element.
snake_case_ :Tuple = 0
while len(required_input[index] ) == 0:
index += 1
if index < len(snake_case ):
snake_case_ :Tuple = required_input[index][0]
if return_tensors is None:
if is_tf_tensor(snake_case ):
snake_case_ :int = """tf"""
elif is_torch_tensor(snake_case ):
snake_case_ :Union[str, Any] = """pt"""
elif isinstance(snake_case , (int, float, list, tuple, np.ndarray) ):
snake_case_ :Optional[int] = """np"""
else:
raise ValueError(
f"""type of {first_element} unknown: {type(snake_case )}. """
"""Should be one of a python, numpy, pytorch or tensorflow object.""" )
for key, value in processed_features.items():
if isinstance(value[0] , (int, float) ):
snake_case_ :Optional[int] = to_numpy(snake_case )
else:
snake_case_ :int = [to_numpy(snake_case ) for v in value]
# Convert padding_strategy in PaddingStrategy
snake_case_ :Union[str, Any] = self._get_padding_strategies(padding=snake_case , max_length=snake_case )
snake_case_ :List[Any] = processed_features[self.model_input_names[0]]
snake_case_ :Optional[Any] = len(snake_case )
if not all(len(snake_case ) == batch_size for v in processed_features.values() ):
raise ValueError("""Some items in the output dictionary have a different batch size than others.""" )
snake_case_ :Optional[Any] = []
for i in range(snake_case ):
snake_case_ :List[Any] = {k: v[i] for k, v in processed_features.items()}
# truncation
snake_case_ :Union[str, Any] = self._truncate(
snake_case , max_length=snake_case , pad_to_multiple_of=snake_case , truncation=snake_case , )
truncated_inputs.append(snake_case )
if padding_strategy == PaddingStrategy.LONGEST:
# make sure that `max_length` cannot be longer than the longest truncated length
snake_case_ :Optional[Any] = max(len(input_slice[self.model_input_names[0]] ) for input_slice in truncated_inputs )
snake_case_ :int = PaddingStrategy.MAX_LENGTH
snake_case_ :List[Any] = {}
for i in range(snake_case ):
# padding
snake_case_ :Any = self._pad(
truncated_inputs[i] , max_length=snake_case , padding_strategy=snake_case , pad_to_multiple_of=snake_case , return_attention_mask=snake_case , )
for key, value in outputs.items():
if key not in batch_outputs:
snake_case_ :Optional[int] = []
if value.dtype is np.dtype(np.floataa ):
snake_case_ :Tuple = value.astype(np.floataa )
batch_outputs[key].append(snake_case )
return BatchFeature(snake_case , tensor_type=snake_case )
def lowerCAmelCase_ ( self: Dict , snake_case: Union[Dict[str, np.ndarray], BatchFeature] , snake_case: Optional[int] = None , snake_case: PaddingStrategy = PaddingStrategy.DO_NOT_PAD , snake_case: Optional[int] = None , snake_case: Optional[bool] = None , ) -> dict:
snake_case_ :Any = processed_features[self.model_input_names[0]]
if padding_strategy == PaddingStrategy.LONGEST:
snake_case_ :Any = len(snake_case )
if max_length is not None and pad_to_multiple_of is not None and (max_length % pad_to_multiple_of != 0):
snake_case_ :Optional[int] = ((max_length // pad_to_multiple_of) + 1) * pad_to_multiple_of
snake_case_ :Tuple = padding_strategy != PaddingStrategy.DO_NOT_PAD and len(snake_case ) < max_length
if return_attention_mask and "attention_mask" not in processed_features:
snake_case_ :Union[str, Any] = np.ones(len(snake_case ) , dtype=np.intaa )
if needs_to_be_padded:
snake_case_ :Optional[int] = max_length - len(snake_case )
if self.padding_side == "right":
if return_attention_mask:
snake_case_ :Union[str, Any] = np.pad(
processed_features["""attention_mask"""] , (0, difference) )
snake_case_ :Tuple = ((0, difference), (0, 0)) if self.feature_size > 1 else (0, difference)
snake_case_ :int = np.pad(
snake_case , snake_case , """constant""" , constant_values=self.padding_value )
elif self.padding_side == "left":
if return_attention_mask:
snake_case_ :Dict = np.pad(
processed_features["""attention_mask"""] , (difference, 0) )
snake_case_ :str = ((difference, 0), (0, 0)) if self.feature_size > 1 else (difference, 0)
snake_case_ :Optional[Any] = np.pad(
snake_case , snake_case , """constant""" , constant_values=self.padding_value )
else:
raise ValueError("""Invalid padding strategy:""" + str(self.padding_side ) )
return processed_features
def lowerCAmelCase_ ( self: Optional[Any] , snake_case: Union[Dict[str, np.ndarray], BatchFeature] , snake_case: Optional[int] = None , snake_case: Optional[int] = None , snake_case: Optional[bool] = None , ) -> List[Any]:
if not truncation:
return processed_features
elif truncation and max_length is None:
raise ValueError("""When setting ``truncation=True``, make sure that ``max_length`` is defined.""" )
snake_case_ :Any = processed_features[self.model_input_names[0]]
# find `max_length` that fits `pad_to_multiple_of`
if max_length is not None and pad_to_multiple_of is not None and (max_length % pad_to_multiple_of != 0):
snake_case_ :str = ((max_length // pad_to_multiple_of) + 1) * pad_to_multiple_of
snake_case_ :Optional[Any] = len(snake_case ) > max_length
if needs_to_be_truncated:
snake_case_ :Union[str, Any] = processed_features[self.model_input_names[0]][:max_length]
if "attention_mask" in processed_features:
snake_case_ :Union[str, Any] = processed_features["""attention_mask"""][:max_length]
return processed_features
def lowerCAmelCase_ ( self: List[Any] , snake_case: Tuple=False , snake_case: Union[str, Any]=None ) -> Union[str, Any]:
# Get padding strategy
if padding is not False:
if padding is True:
snake_case_ :Optional[Any] = PaddingStrategy.LONGEST # Default to pad to the longest sequence in the batch
elif not isinstance(snake_case , snake_case ):
snake_case_ :Dict = PaddingStrategy(snake_case )
elif isinstance(snake_case , snake_case ):
snake_case_ :Optional[Any] = padding
else:
snake_case_ :Optional[Any] = PaddingStrategy.DO_NOT_PAD
# Set max length if needed
if max_length is None:
if padding_strategy == PaddingStrategy.MAX_LENGTH:
raise ValueError(
f"""When setting ``padding={PaddingStrategy.MAX_LENGTH}``, make sure that max_length is defined""" )
# Test if we have a padding value
if padding_strategy != PaddingStrategy.DO_NOT_PAD and (self.padding_value is None):
raise ValueError(
"""Asking to pad but the feature_extractor does not have a padding value. Please select a value to use"""
""" as `padding_value`. For example: `feature_extractor.padding_value = 0.0`.""" )
return padding_strategy
| 66 |
"""simple docstring"""
import collections
import inspect
import unittest
from transformers import SwinvaConfig
from transformers.testing_utils import require_torch, require_vision, slow, torch_device
from transformers.utils import cached_property, is_torch_available, is_vision_available
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, _config_zero_init, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from torch import nn
from transformers import SwinvaForImageClassification, SwinvaForMaskedImageModeling, SwinvaModel
from transformers.models.swinva.modeling_swinva import SWINV2_PRETRAINED_MODEL_ARCHIVE_LIST
if is_vision_available():
from PIL import Image
from transformers import AutoImageProcessor
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Optional[int] , snake_case: Any , snake_case: Optional[Any]=13 , snake_case: Tuple=32 , snake_case: Optional[int]=2 , snake_case: Tuple=3 , snake_case: Tuple=16 , snake_case: Optional[Any]=[1, 2, 1] , snake_case: Optional[int]=[2, 2, 4] , snake_case: Optional[int]=2 , snake_case: int=2.0 , snake_case: Union[str, Any]=True , snake_case: List[str]=0.0 , snake_case: List[Any]=0.0 , snake_case: Optional[Any]=0.1 , snake_case: List[Any]="gelu" , snake_case: Optional[int]=False , snake_case: Union[str, Any]=True , snake_case: Union[str, Any]=0.0_2 , snake_case: Optional[int]=1E-5 , snake_case: Optional[Any]=True , snake_case: List[Any]=None , snake_case: List[Any]=True , snake_case: Optional[Any]=10 , snake_case: str=8 , ) -> Tuple:
snake_case_ :Dict = parent
snake_case_ :Any = batch_size
snake_case_ :List[Any] = image_size
snake_case_ :List[Any] = patch_size
snake_case_ :int = num_channels
snake_case_ :Tuple = embed_dim
snake_case_ :str = depths
snake_case_ :str = num_heads
snake_case_ :Optional[int] = window_size
snake_case_ :Tuple = mlp_ratio
snake_case_ :Any = qkv_bias
snake_case_ :List[Any] = hidden_dropout_prob
snake_case_ :Optional[Any] = attention_probs_dropout_prob
snake_case_ :Union[str, Any] = drop_path_rate
snake_case_ :Any = hidden_act
snake_case_ :Optional[Any] = use_absolute_embeddings
snake_case_ :Union[str, Any] = patch_norm
snake_case_ :Dict = layer_norm_eps
snake_case_ :str = initializer_range
snake_case_ :Tuple = is_training
snake_case_ :Tuple = scope
snake_case_ :Union[str, Any] = use_labels
snake_case_ :Optional[Any] = type_sequence_label_size
snake_case_ :Dict = encoder_stride
def lowerCAmelCase_ ( self: int ) -> int:
snake_case_ :List[str] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case_ :Any = None
if self.use_labels:
snake_case_ :str = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ :int = self.get_config()
return config, pixel_values, labels
def lowerCAmelCase_ ( self: str ) -> Union[str, Any]:
return SwinvaConfig(
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 , )
def lowerCAmelCase_ ( self: str , snake_case: Optional[int] , snake_case: Dict , snake_case: str ) -> List[Any]:
snake_case_ :Union[str, Any] = SwinvaModel(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Optional[int] = model(snake_case )
snake_case_ :Optional[Any] = ((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 lowerCAmelCase_ ( self: int , snake_case: List[str] , snake_case: Tuple , snake_case: int ) -> Any:
snake_case_ :Dict = SwinvaForMaskedImageModeling(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Tuple = model(snake_case )
self.parent.assertEqual(
result.logits.shape , (self.batch_size, self.num_channels, self.image_size, self.image_size) )
# test greyscale images
snake_case_ :List[Any] = 1
snake_case_ :int = SwinvaForMaskedImageModeling(snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Union[str, Any] = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] )
snake_case_ :int = model(snake_case )
self.parent.assertEqual(result.logits.shape , (self.batch_size, 1, self.image_size, self.image_size) )
def lowerCAmelCase_ ( self: List[Any] , snake_case: Any , snake_case: List[str] , snake_case: Union[str, Any] ) -> Tuple:
snake_case_ :int = self.type_sequence_label_size
snake_case_ :List[Any] = SwinvaForImageClassification(snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Dict = model(snake_case , labels=snake_case )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
def lowerCAmelCase_ ( self: int ) -> str:
snake_case_ :Any = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_ :List[str] = config_and_inputs
snake_case_ :List[Any] = {"""pixel_values""": pixel_values}
return config, inputs_dict
@require_torch
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : Optional[Any] = (
(SwinvaModel, SwinvaForImageClassification, SwinvaForMaskedImageModeling) if is_torch_available() else ()
)
_A : Any = (
{"""feature-extraction""": SwinvaModel, """image-classification""": SwinvaForImageClassification}
if is_torch_available()
else {}
)
_A : List[Any] = False
_A : List[str] = False
_A : Tuple = False
_A : List[str] = False
def lowerCAmelCase_ ( self: Dict ) -> List[Any]:
snake_case_ :Optional[int] = SwinvaModelTester(self )
snake_case_ :List[str] = ConfigTester(self , config_class=snake_case , embed_dim=37 )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> List[Any]:
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 lowerCAmelCase_ ( self: Union[str, Any] ) -> Tuple:
snake_case_ :List[str] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case )
@unittest.skip(reason="""Got `CUDA error: misaligned address` with PyTorch 2.0.0.""" )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> str:
pass
@unittest.skip(reason="""Swinv2 does not use inputs_embeds""" )
def lowerCAmelCase_ ( self: int ) -> Dict:
pass
def lowerCAmelCase_ ( self: List[str] ) -> Union[str, Any]:
snake_case_, snake_case_ :List[str] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :Optional[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 lowerCAmelCase_ ( self: Dict ) -> Optional[int]:
snake_case_, snake_case_ :Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :Optional[int] = model_class(snake_case )
snake_case_ :List[Any] = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case_ :int = [*signature.parameters.keys()]
snake_case_ :List[Any] = ["""pixel_values"""]
self.assertListEqual(arg_names[:1] , snake_case )
def lowerCAmelCase_ ( self: List[str] ) -> Optional[Any]:
snake_case_, snake_case_ :List[str] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :List[str] = True
for model_class in self.all_model_classes:
snake_case_ :List[Any] = True
snake_case_ :Any = False
snake_case_ :Optional[int] = True
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_ :str = outputs.attentions
snake_case_ :Dict = len(self.model_tester.depths )
self.assertEqual(len(snake_case ) , snake_case )
# check that output_attentions also work using config
del inputs_dict["output_attentions"]
snake_case_ :Union[str, Any] = True
snake_case_ :Tuple = config.window_size**2
snake_case_ :Any = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :Union[str, Any] = model(**self._prepare_for_class(snake_case , snake_case ) )
snake_case_ :int = outputs.attentions
self.assertEqual(len(snake_case ) , snake_case )
self.assertListEqual(
list(attentions[0].shape[-3:] ) , [self.model_tester.num_heads[0], window_size_squared, window_size_squared] , )
snake_case_ :Any = len(snake_case )
# Check attention is always last and order is fine
snake_case_ :int = True
snake_case_ :Dict = True
snake_case_ :Optional[int] = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :Dict = model(**self._prepare_for_class(snake_case , snake_case ) )
if hasattr(self.model_tester , """num_hidden_states_types""" ):
snake_case_ :Any = self.model_tester.num_hidden_states_types
else:
# also another +1 for reshaped_hidden_states
snake_case_ :int = 2
self.assertEqual(out_len + added_hidden_states , len(snake_case ) )
snake_case_ :str = outputs.attentions
self.assertEqual(len(snake_case ) , snake_case )
self.assertListEqual(
list(self_attentions[0].shape[-3:] ) , [self.model_tester.num_heads[0], window_size_squared, window_size_squared] , )
def lowerCAmelCase_ ( self: int , snake_case: Dict , snake_case: Dict , snake_case: Optional[Any] , snake_case: Dict ) -> List[str]:
snake_case_ :Dict = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :Optional[int] = model(**self._prepare_for_class(snake_case , snake_case ) )
snake_case_ :str = outputs.hidden_states
snake_case_ :List[Any] = getattr(
self.model_tester , """expected_num_hidden_layers""" , len(self.model_tester.depths ) + 1 )
self.assertEqual(len(snake_case ) , snake_case )
# Swinv2 has a different seq_length
snake_case_ :List[Any] = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :Optional[int] = (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] , )
snake_case_ :str = outputs.reshaped_hidden_states
self.assertEqual(len(snake_case ) , snake_case )
snake_case_, snake_case_, snake_case_, snake_case_ :Any = reshaped_hidden_states[0].shape
snake_case_ :int = (
reshaped_hidden_states[0].view(snake_case , snake_case , height * width ).permute(0 , 2 , 1 )
)
self.assertListEqual(
list(reshaped_hidden_states.shape[-2:] ) , [num_patches, self.model_tester.embed_dim] , )
def lowerCAmelCase_ ( self: Any ) -> Any:
snake_case_, snake_case_ :List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Union[str, Any] = (
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_ :Union[str, Any] = 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_ :List[str] = True
self.check_hidden_states_output(snake_case , snake_case , snake_case , snake_case )
def lowerCAmelCase_ ( self: Tuple ) -> Any:
snake_case_, snake_case_ :Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Optional[int] = 3
snake_case_ :Union[str, Any] = (
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_ :str = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :Any = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
snake_case_ :int = 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_ :Tuple = True
self.check_hidden_states_output(snake_case , snake_case , snake_case , (padded_height, padded_width) )
def lowerCAmelCase_ ( self: Any ) -> Tuple:
snake_case_ :int = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_masked_image_modeling(*snake_case )
def lowerCAmelCase_ ( self: Optional[int] ) -> Dict:
snake_case_ :Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_classification(*snake_case )
@slow
def lowerCAmelCase_ ( self: List[Any] ) -> Dict:
for model_name in SWINV2_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
snake_case_ :List[str] = SwinvaModel.from_pretrained(snake_case )
self.assertIsNotNone(snake_case )
def lowerCAmelCase_ ( self: Optional[int] ) -> List[Any]:
snake_case_, snake_case_ :str = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Optional[int] = _config_zero_init(snake_case )
for model_class in self.all_model_classes:
snake_case_ :Tuple = model_class(config=snake_case )
for name, param in model.named_parameters():
if "embeddings" not in name and "logit_scale" not in name and param.requires_grad:
self.assertIn(
((param.data.mean() * 1E9).round() / 1E9).item() , [0.0, 1.0] , msg=f"""Parameter {name} of model {model_class} seems not properly initialized""" , )
@require_vision
@require_torch
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
@cached_property
def lowerCAmelCase_ ( self: Optional[int] ) -> List[Any]:
return (
AutoImageProcessor.from_pretrained("""microsoft/swinv2-tiny-patch4-window8-256""" )
if is_vision_available()
else None
)
@slow
def lowerCAmelCase_ ( self: List[str] ) -> List[str]:
snake_case_ :Tuple = SwinvaForImageClassification.from_pretrained("""microsoft/swinv2-tiny-patch4-window8-256""" ).to(
snake_case )
snake_case_ :str = self.default_image_processor
snake_case_ :List[Any] = Image.open("""./tests/fixtures/tests_samples/COCO/000000039769.png""" )
snake_case_ :str = image_processor(images=snake_case , return_tensors="""pt""" ).to(snake_case )
# forward pass
with torch.no_grad():
snake_case_ :Tuple = model(**snake_case )
# verify the logits
snake_case_ :Dict = torch.Size((1, 1_000) )
self.assertEqual(outputs.logits.shape , snake_case )
snake_case_ :int = torch.tensor([-0.3_9_4_7, -0.4_3_0_6, 0.0_0_2_6] ).to(snake_case )
self.assertTrue(torch.allclose(outputs.logits[0, :3] , snake_case , atol=1E-4 ) )
| 66 | 1 |
"""simple docstring"""
import os
import pytest
import yaml
from datasets.features.features import Features, Value
from datasets.info import DatasetInfo, DatasetInfosDict
@pytest.mark.parametrize(
"""files""", [
["""full:README.md""", """dataset_infos.json"""],
["""empty:README.md""", """dataset_infos.json"""],
["""dataset_infos.json"""],
["""full:README.md"""],
], )
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :Union[str, Any] = tmp_path_factory.mktemp("""dset_infos_dir""" )
if "full:README.md" in files:
with open(dataset_infos_dir / """README.md""", """w""" ) as f:
f.write("""---\ndataset_info:\n dataset_size: 42\n---""" )
if "empty:README.md" in files:
with open(dataset_infos_dir / """README.md""", """w""" ) as f:
f.write("""""" )
# we want to support dataset_infos.json for backward compatibility
if "dataset_infos.json" in files:
with open(dataset_infos_dir / """dataset_infos.json""", """w""" ) as f:
f.write("""{\"default\": {\"dataset_size\": 42}}""" )
snake_case_ :Any = DatasetInfosDict.from_directory(_lowercase )
assert dataset_infos
assert dataset_infos["default"].dataset_size == 42
@pytest.mark.parametrize(
"""dataset_info""", [
DatasetInfo(),
DatasetInfo(
description="""foo""", features=Features({"""a""": Value("""int32""" )} ), builder_name="""builder""", config_name="""config""", version="""1.0.0""", splits=[{"""name""": """train"""}], download_size=42, ),
], )
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :int = str(_lowercase )
dataset_info.write_to_directory(_lowercase )
snake_case_ :List[Any] = DatasetInfo.from_directory(_lowercase )
assert dataset_info == reloaded
assert os.path.exists(os.path.join(_lowercase, """dataset_info.json""" ) )
def A_ ( ):
'''simple docstring'''
snake_case_ :int = DatasetInfo(
description="""foo""", citation="""bar""", homepage="""https://foo.bar""", license="""CC0""", features=Features({"""a""": Value("""int32""" )} ), post_processed={}, supervised_keys=(), task_templates=[], builder_name="""builder""", config_name="""config""", version="""1.0.0""", splits=[{"""name""": """train""", """num_examples""": 42}], download_checksums={}, download_size=1337, post_processing_size=442, dataset_size=1234, size_in_bytes=1337 + 442 + 1234, )
snake_case_ :List[Any] = dataset_info._to_yaml_dict()
assert sorted(_lowercase ) == sorted(DatasetInfo._INCLUDED_INFO_IN_YAML )
for key in DatasetInfo._INCLUDED_INFO_IN_YAML:
assert key in dataset_info_yaml_dict
assert isinstance(dataset_info_yaml_dict[key], (list, dict, int, str) )
snake_case_ :Any = yaml.safe_dump(_lowercase )
snake_case_ :Optional[Any] = yaml.safe_load(_lowercase )
assert dataset_info_yaml_dict == reloaded
def A_ ( ):
'''simple docstring'''
snake_case_ :int = DatasetInfo()
snake_case_ :Optional[Any] = dataset_info._to_yaml_dict()
assert dataset_info_yaml_dict == {}
@pytest.mark.parametrize(
"""dataset_infos_dict""", [
DatasetInfosDict(),
DatasetInfosDict({"""default""": DatasetInfo()} ),
DatasetInfosDict({"""my_config_name""": DatasetInfo()} ),
DatasetInfosDict(
{
"""default""": DatasetInfo(
description="""foo""", features=Features({"""a""": Value("""int32""" )} ), builder_name="""builder""", config_name="""config""", version="""1.0.0""", splits=[{"""name""": """train"""}], download_size=42, )
} ),
DatasetInfosDict(
{
"""v1""": DatasetInfo(dataset_size=42 ),
"""v2""": DatasetInfo(dataset_size=1337 ),
} ),
], )
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :str = str(_lowercase )
dataset_infos_dict.write_to_directory(_lowercase )
snake_case_ :Dict = DatasetInfosDict.from_directory(_lowercase )
# the config_name of the dataset_infos_dict take over the attribute
for config_name, dataset_info in dataset_infos_dict.items():
snake_case_ :Tuple = config_name
# the yaml representation doesn't include fields like description or citation
# so we just test that we can recover what we can from the yaml
snake_case_ :List[str] = DatasetInfo._from_yaml_dict(dataset_info._to_yaml_dict() )
assert dataset_infos_dict == reloaded
if dataset_infos_dict:
assert os.path.exists(os.path.join(_lowercase, """README.md""" ) )
| 66 |
"""simple docstring"""
import re
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Optional[int] = re.compile(
r"""^(?:0|94|\+94|0{2}94)""" r"""7(0|1|2|4|5|6|7|8)""" r"""(-| |)""" r"""\d{7}$""" )
return bool(re.search(_lowercase, _lowercase ) )
if __name__ == "__main__":
__a = "0094702343221"
print(is_sri_lankan_phone_number(phone))
| 66 | 1 |
"""simple docstring"""
import json
import os
import unittest
from transformers.models.blenderbot_small.tokenization_blenderbot_small import (
VOCAB_FILES_NAMES,
BlenderbotSmallTokenizer,
)
from ...test_tokenization_common import TokenizerTesterMixin
class lowerCamelCase ( _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : Union[str, Any] = BlenderbotSmallTokenizer
_A : List[Any] = False
def lowerCAmelCase_ ( self: Any ) -> Dict:
super().setUp()
snake_case_ :Dict = ["""__start__""", """adapt""", """act""", """ap@@""", """te""", """__end__""", """__unk__"""]
snake_case_ :Dict = dict(zip(snake_case , range(len(snake_case ) ) ) )
snake_case_ :Union[str, Any] = ["""#version: 0.2""", """a p""", """t e</w>""", """ap t</w>""", """a d""", """ad apt</w>""", """a c""", """ac t</w>""", """"""]
snake_case_ :Dict = {"""unk_token""": """__unk__""", """bos_token""": """__start__""", """eos_token""": """__end__"""}
snake_case_ :Any = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES["""vocab_file"""] )
snake_case_ :str = 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 lowerCAmelCase_ ( self: List[Any] , **snake_case: Optional[int] ) -> str:
kwargs.update(self.special_tokens_map )
return BlenderbotSmallTokenizer.from_pretrained(self.tmpdirname , **snake_case )
def lowerCAmelCase_ ( self: Any , snake_case: Optional[Any] ) -> Any:
snake_case_ :Any = """adapt act apte"""
snake_case_ :int = """adapt act apte"""
return input_text, output_text
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
snake_case_ :Any = BlenderbotSmallTokenizer(self.vocab_file , self.merges_file , **self.special_tokens_map )
snake_case_ :Dict = """adapt act apte"""
snake_case_ :Optional[Any] = ["""adapt""", """act""", """ap@@""", """te"""]
snake_case_ :List[str] = tokenizer.tokenize(snake_case )
self.assertListEqual(snake_case , snake_case )
snake_case_ :int = [tokenizer.bos_token] + tokens + [tokenizer.eos_token]
snake_case_ :str = [0, 1, 2, 3, 4, 5]
self.assertListEqual(tokenizer.convert_tokens_to_ids(snake_case ) , snake_case )
def lowerCAmelCase_ ( self: Optional[Any] ) -> Union[str, Any]:
snake_case_ :List[Any] = BlenderbotSmallTokenizer.from_pretrained("""facebook/blenderbot-90M""" )
assert tok("""sam""" ).input_ids == [1_384]
snake_case_ :str = """I am a small frog."""
snake_case_ :Dict = tok([src_text] , padding=snake_case , truncation=snake_case )["""input_ids"""]
snake_case_ :Optional[Any] = tok.batch_decode(snake_case , skip_special_tokens=snake_case , clean_up_tokenization_spaces=snake_case )[0]
assert src_text != decoded # I wish it did!
assert decoded == "i am a small frog ."
def lowerCAmelCase_ ( self: str ) -> str:
snake_case_ :Optional[int] = BlenderbotSmallTokenizer.from_pretrained("""facebook/blenderbot-90M""" )
snake_case_ :str = """I am a small frog ."""
snake_case_ :Dict = """."""
snake_case_ :int = tok(snake_case )["""input_ids"""]
snake_case_ :List[Any] = tok(snake_case )["""input_ids"""]
assert encoded[-1] == encoded_dot[0]
| 66 |
"""simple docstring"""
import argparse
import json
import os
import pickle
import shutil
import numpy as np
import torch
from distiller import Distiller
from lm_seqs_dataset import LmSeqsDataset
from transformers import (
BertConfig,
BertForMaskedLM,
BertTokenizer,
DistilBertConfig,
DistilBertForMaskedLM,
DistilBertTokenizer,
GPTaConfig,
GPTaLMHeadModel,
GPTaTokenizer,
RobertaConfig,
RobertaForMaskedLM,
RobertaTokenizer,
)
from utils import git_log, init_gpu_params, logger, set_seed
__a = {
"distilbert": (DistilBertConfig, DistilBertForMaskedLM, DistilBertTokenizer),
"roberta": (RobertaConfig, RobertaForMaskedLM, RobertaTokenizer),
"bert": (BertConfig, BertForMaskedLM, BertTokenizer),
"gpt2": (GPTaConfig, GPTaLMHeadModel, GPTaTokenizer),
}
def A_ ( _lowercase ):
'''simple docstring'''
assert (args.mlm and args.alpha_mlm > 0.0) or (not args.mlm and args.alpha_mlm == 0.0)
assert (args.alpha_mlm > 0.0 and args.alpha_clm == 0.0) or (args.alpha_mlm == 0.0 and args.alpha_clm > 0.0)
if args.mlm:
assert os.path.isfile(args.token_counts )
assert (args.student_type in ["roberta", "distilbert"]) and (args.teacher_type in ["roberta", "bert"])
else:
assert (args.student_type in ["gpt2"]) and (args.teacher_type in ["gpt2"])
assert args.teacher_type == args.student_type or (
args.student_type == "distilbert" and args.teacher_type == "bert"
)
assert os.path.isfile(args.student_config )
if args.student_pretrained_weights is not None:
assert os.path.isfile(args.student_pretrained_weights )
if args.freeze_token_type_embds:
assert args.student_type in ["roberta"]
assert args.alpha_ce >= 0.0
assert args.alpha_mlm >= 0.0
assert args.alpha_clm >= 0.0
assert args.alpha_mse >= 0.0
assert args.alpha_cos >= 0.0
assert args.alpha_ce + args.alpha_mlm + args.alpha_clm + args.alpha_mse + args.alpha_cos > 0.0
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
if args.student_type == "roberta":
snake_case_ :Tuple = False
elif args.student_type == "gpt2":
snake_case_ :Union[str, Any] = False
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
if args.student_type == "roberta":
snake_case_ :List[str] = False
def A_ ( ):
'''simple docstring'''
snake_case_ :Union[str, Any] = argparse.ArgumentParser(description="""Training""" )
parser.add_argument("""--force""", action="""store_true""", help="""Overwrite dump_path if it already exists.""" )
parser.add_argument(
"""--dump_path""", type=_lowercase, required=_lowercase, help="""The output directory (log, checkpoints, parameters, etc.)""" )
parser.add_argument(
"""--data_file""", type=_lowercase, required=_lowercase, help="""The binarized file (tokenized + tokens_to_ids) and grouped by sequence.""", )
parser.add_argument(
"""--student_type""", type=_lowercase, choices=["""distilbert""", """roberta""", """gpt2"""], required=_lowercase, help="""The student type (DistilBERT, RoBERTa).""", )
parser.add_argument("""--student_config""", type=_lowercase, required=_lowercase, help="""Path to the student configuration.""" )
parser.add_argument(
"""--student_pretrained_weights""", default=_lowercase, type=_lowercase, help="""Load student initialization checkpoint.""" )
parser.add_argument(
"""--teacher_type""", choices=["""bert""", """roberta""", """gpt2"""], required=_lowercase, help="""Teacher type (BERT, RoBERTa).""" )
parser.add_argument("""--teacher_name""", type=_lowercase, required=_lowercase, help="""The teacher model.""" )
parser.add_argument("""--temperature""", default=2.0, type=_lowercase, help="""Temperature for the softmax temperature.""" )
parser.add_argument(
"""--alpha_ce""", default=0.5, type=_lowercase, help="""Linear weight for the distillation loss. Must be >=0.""" )
parser.add_argument(
"""--alpha_mlm""", default=0.0, type=_lowercase, help="""Linear weight for the MLM loss. Must be >=0. Should be used in conjunction with `mlm` flag.""", )
parser.add_argument("""--alpha_clm""", default=0.5, type=_lowercase, help="""Linear weight for the CLM loss. Must be >=0.""" )
parser.add_argument("""--alpha_mse""", default=0.0, type=_lowercase, help="""Linear weight of the MSE loss. Must be >=0.""" )
parser.add_argument(
"""--alpha_cos""", default=0.0, type=_lowercase, help="""Linear weight of the cosine embedding loss. Must be >=0.""" )
parser.add_argument(
"""--mlm""", action="""store_true""", help="""The LM step: MLM or CLM. If `mlm` is True, the MLM is used over CLM.""" )
parser.add_argument(
"""--mlm_mask_prop""", default=0.15, type=_lowercase, help="""Proportion of tokens for which we need to make a prediction.""", )
parser.add_argument("""--word_mask""", default=0.8, type=_lowercase, help="""Proportion of tokens to mask out.""" )
parser.add_argument("""--word_keep""", default=0.1, type=_lowercase, help="""Proportion of tokens to keep.""" )
parser.add_argument("""--word_rand""", default=0.1, type=_lowercase, help="""Proportion of tokens to randomly replace.""" )
parser.add_argument(
"""--mlm_smoothing""", default=0.7, type=_lowercase, help="""Smoothing parameter to emphasize more rare tokens (see XLM, similar to word2vec).""", )
parser.add_argument("""--token_counts""", type=_lowercase, help="""The token counts in the data_file for MLM.""" )
parser.add_argument(
"""--restrict_ce_to_mask""", action="""store_true""", help="""If true, compute the distillation loss only the [MLM] prediction distribution.""", )
parser.add_argument(
"""--freeze_pos_embs""", action="""store_true""", help="""Freeze positional embeddings during distillation. For student_type in ['roberta', 'gpt2'] only.""", )
parser.add_argument(
"""--freeze_token_type_embds""", action="""store_true""", help="""Freeze token type embeddings during distillation if existent. For student_type in ['roberta'] only.""", )
parser.add_argument("""--n_epoch""", type=_lowercase, default=3, help="""Number of pass on the whole dataset.""" )
parser.add_argument("""--batch_size""", type=_lowercase, default=5, help="""Batch size (for each process).""" )
parser.add_argument(
"""--group_by_size""", action="""store_false""", help="""If true, group sequences that have similar length into the same batch. Default is true.""", )
parser.add_argument(
"""--gradient_accumulation_steps""", type=_lowercase, default=50, help="""Gradient accumulation for larger training batches.""", )
parser.add_argument("""--warmup_prop""", default=0.05, type=_lowercase, help="""Linear warmup proportion.""" )
parser.add_argument("""--weight_decay""", default=0.0, type=_lowercase, help="""Weight decay if we apply some.""" )
parser.add_argument("""--learning_rate""", default=5e-4, type=_lowercase, help="""The initial learning rate for Adam.""" )
parser.add_argument("""--adam_epsilon""", default=1e-6, type=_lowercase, help="""Epsilon for Adam optimizer.""" )
parser.add_argument("""--max_grad_norm""", default=5.0, type=_lowercase, help="""Max gradient norm.""" )
parser.add_argument("""--initializer_range""", default=0.02, type=_lowercase, help="""Random initialization range.""" )
parser.add_argument(
"""--fp16""", action="""store_true""", help="""Whether to use 16-bit (mixed) precision (through NVIDIA apex) instead of 32-bit""", )
parser.add_argument(
"""--fp16_opt_level""", type=_lowercase, default="""O1""", help=(
"""For fp16: Apex AMP optimization level selected in ['O0', 'O1', 'O2', and 'O3']."""
"""See details at https://nvidia.github.io/apex/amp.html"""
), )
parser.add_argument("""--n_gpu""", type=_lowercase, default=1, help="""Number of GPUs in the node.""" )
parser.add_argument("""--local_rank""", type=_lowercase, default=-1, help="""Distributed training - Local rank""" )
parser.add_argument("""--seed""", type=_lowercase, default=56, help="""Random seed""" )
parser.add_argument("""--log_interval""", type=_lowercase, default=500, help="""Tensorboard logging interval.""" )
parser.add_argument("""--checkpoint_interval""", type=_lowercase, default=4000, help="""Checkpoint interval.""" )
snake_case_ :Tuple = parser.parse_args()
sanity_checks(_lowercase )
# ARGS #
init_gpu_params(_lowercase )
set_seed(_lowercase )
if args.is_master:
if os.path.exists(args.dump_path ):
if not args.force:
raise ValueError(
f"""Serialization dir {args.dump_path} already exists, but you have not precised wheter to overwrite"""
""" itUse `--force` if you want to overwrite it""" )
else:
shutil.rmtree(args.dump_path )
if not os.path.exists(args.dump_path ):
os.makedirs(args.dump_path )
logger.info(f"""Experiment will be dumped and logged in {args.dump_path}""" )
# SAVE PARAMS #
logger.info(f"""Param: {args}""" )
with open(os.path.join(args.dump_path, """parameters.json""" ), """w""" ) as f:
json.dump(vars(_lowercase ), _lowercase, indent=4 )
git_log(args.dump_path )
snake_case_, snake_case_, snake_case_ :Any = MODEL_CLASSES[args.student_type]
snake_case_, snake_case_, snake_case_ :int = MODEL_CLASSES[args.teacher_type]
# TOKENIZER #
snake_case_ :Any = teacher_tokenizer_class.from_pretrained(args.teacher_name )
snake_case_ :Optional[Any] = {}
for tok_name, tok_symbol in tokenizer.special_tokens_map.items():
snake_case_ :Union[str, Any] = tokenizer.all_special_tokens.index(_lowercase )
snake_case_ :Union[str, Any] = tokenizer.all_special_ids[idx]
logger.info(f"""Special tokens {special_tok_ids}""" )
snake_case_ :str = special_tok_ids
snake_case_ :Any = tokenizer.max_model_input_sizes[args.teacher_name]
# DATA LOADER #
logger.info(f"""Loading data from {args.data_file}""" )
with open(args.data_file, """rb""" ) as fp:
snake_case_ :str = pickle.load(_lowercase )
if args.mlm:
logger.info(f"""Loading token counts from {args.token_counts} (already pre-computed)""" )
with open(args.token_counts, """rb""" ) as fp:
snake_case_ :Optional[Any] = pickle.load(_lowercase )
snake_case_ :Tuple = np.maximum(_lowercase, 1 ) ** -args.mlm_smoothing
for idx in special_tok_ids.values():
snake_case_ :Optional[int] = 0.0 # do not predict special tokens
snake_case_ :int = torch.from_numpy(_lowercase )
else:
snake_case_ :List[str] = None
snake_case_ :Optional[int] = LmSeqsDataset(params=_lowercase, data=_lowercase )
logger.info("""Data loader created.""" )
# STUDENT #
logger.info(f"""Loading student config from {args.student_config}""" )
snake_case_ :List[Any] = student_config_class.from_pretrained(args.student_config )
snake_case_ :Union[str, Any] = True
if args.student_pretrained_weights is not None:
logger.info(f"""Loading pretrained weights from {args.student_pretrained_weights}""" )
snake_case_ :List[str] = student_model_class.from_pretrained(args.student_pretrained_weights, config=_lowercase )
else:
snake_case_ :Optional[int] = student_model_class(_lowercase )
if args.n_gpu > 0:
student.to(f"""cuda:{args.local_rank}""" )
logger.info("""Student loaded.""" )
# TEACHER #
snake_case_ :Dict = teacher_model_class.from_pretrained(args.teacher_name, output_hidden_states=_lowercase )
if args.n_gpu > 0:
teacher.to(f"""cuda:{args.local_rank}""" )
logger.info(f"""Teacher loaded from {args.teacher_name}.""" )
# FREEZING #
if args.freeze_pos_embs:
freeze_pos_embeddings(_lowercase, _lowercase )
if args.freeze_token_type_embds:
freeze_token_type_embeddings(_lowercase, _lowercase )
# SANITY CHECKS #
assert student.config.vocab_size == teacher.config.vocab_size
assert student.config.hidden_size == teacher.config.hidden_size
assert student.config.max_position_embeddings == teacher.config.max_position_embeddings
if args.mlm:
assert token_probs.size(0 ) == stu_architecture_config.vocab_size
# DISTILLER #
torch.cuda.empty_cache()
snake_case_ :Optional[int] = Distiller(
params=_lowercase, dataset=_lowercase, token_probs=_lowercase, student=_lowercase, teacher=_lowercase )
distiller.train()
logger.info("""Let's go get some drinks.""" )
if __name__ == "__main__":
main()
| 66 | 1 |
"""simple docstring"""
from typing import Any, Dict, List, Union
from ..utils import add_end_docstrings, is_torch_available, is_vision_available, logging, requires_backends
from .base import PIPELINE_INIT_ARGS, ChunkPipeline
if is_vision_available():
from PIL import Image
from ..image_utils import load_image
if is_torch_available():
import torch
from transformers.modeling_outputs import BaseModelOutput
from ..models.auto.modeling_auto import MODEL_FOR_ZERO_SHOT_OBJECT_DETECTION_MAPPING
__a = logging.get_logger(__name__)
@add_end_docstrings(_lowerCAmelCase )
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
def __init__( self: Union[str, Any] , **snake_case: int ) -> Any:
super().__init__(**snake_case )
if self.framework == "tf":
raise ValueError(f"""The {self.__class__} is only available in PyTorch.""" )
requires_backends(self , """vision""" )
self.check_model_type(snake_case )
def __call__( self: Union[str, Any] , snake_case: Union[str, "Image.Image", List[Dict[str, Any]]] , snake_case: Union[str, List[str]] = None , **snake_case: Union[str, Any] , ) -> Dict:
if "text_queries" in kwargs:
snake_case_ :Optional[Any] = kwargs.pop("""text_queries""" )
if isinstance(snake_case , (str, Image.Image) ):
snake_case_ :Tuple = {"""image""": image, """candidate_labels""": candidate_labels}
else:
snake_case_ :Dict = image
snake_case_ :Optional[Any] = super().__call__(snake_case , **snake_case )
return results
def lowerCAmelCase_ ( self: Any , **snake_case: Dict ) -> Tuple:
snake_case_ :int = {}
if "threshold" in kwargs:
snake_case_ :Optional[Any] = kwargs["""threshold"""]
if "top_k" in kwargs:
snake_case_ :Any = kwargs["""top_k"""]
return {}, {}, postprocess_params
def lowerCAmelCase_ ( self: Union[str, Any] , snake_case: Any ) -> str:
snake_case_ :List[str] = load_image(inputs["""image"""] )
snake_case_ :int = inputs["""candidate_labels"""]
if isinstance(snake_case , snake_case ):
snake_case_ :Union[str, Any] = candidate_labels.split(""",""" )
snake_case_ :List[str] = torch.tensor([[image.height, image.width]] , dtype=torch.intaa )
for i, candidate_label in enumerate(snake_case ):
snake_case_ :Dict = self.tokenizer(snake_case , return_tensors=self.framework )
snake_case_ :Optional[int] = self.image_processor(snake_case , return_tensors=self.framework )
yield {
"is_last": i == len(snake_case ) - 1,
"target_size": target_size,
"candidate_label": candidate_label,
**text_inputs,
**image_features,
}
def lowerCAmelCase_ ( self: Optional[Any] , snake_case: Optional[int] ) -> Optional[Any]:
snake_case_ :Union[str, Any] = model_inputs.pop("""target_size""" )
snake_case_ :Any = model_inputs.pop("""candidate_label""" )
snake_case_ :Optional[int] = model_inputs.pop("""is_last""" )
snake_case_ :Optional[Any] = self.model(**snake_case )
snake_case_ :Union[str, Any] = {"""target_size""": target_size, """candidate_label""": candidate_label, """is_last""": is_last, **outputs}
return model_outputs
def lowerCAmelCase_ ( self: List[Any] , snake_case: List[str] , snake_case: Optional[int]=0.1 , snake_case: Optional[Any]=None ) -> Tuple:
snake_case_ :List[Any] = []
for model_output in model_outputs:
snake_case_ :List[str] = model_output["""candidate_label"""]
snake_case_ :Any = BaseModelOutput(snake_case )
snake_case_ :int = self.image_processor.post_process_object_detection(
outputs=snake_case , threshold=snake_case , target_sizes=model_output["""target_size"""] )[0]
for index in outputs["scores"].nonzero():
snake_case_ :Union[str, Any] = outputs["""scores"""][index].item()
snake_case_ :Optional[int] = self._get_bounding_box(outputs["""boxes"""][index][0] )
snake_case_ :Dict = {"""score""": score, """label""": label, """box""": box}
results.append(snake_case )
snake_case_ :List[str] = sorted(snake_case , key=lambda snake_case : x["score"] , reverse=snake_case )
if top_k:
snake_case_ :Any = results[:top_k]
return results
def lowerCAmelCase_ ( self: Optional[int] , snake_case: "torch.Tensor" ) -> Dict[str, int]:
if self.framework != "pt":
raise ValueError("""The ZeroShotObjectDetectionPipeline is only available in PyTorch.""" )
snake_case_, snake_case_, snake_case_, snake_case_ :Dict = box.int().tolist()
snake_case_ :Any = {
"""xmin""": xmin,
"""ymin""": ymin,
"""xmax""": xmax,
"""ymax""": ymax,
}
return bbox
| 66 |
"""simple docstring"""
import json
import os
import subprocess
import unittest
from ast import literal_eval
import pytest
from parameterized import parameterized, parameterized_class
from . import is_sagemaker_available
if is_sagemaker_available():
from sagemaker import Session, TrainingJobAnalytics
from sagemaker.huggingface import HuggingFace
@pytest.mark.skipif(
literal_eval(os.getenv("""TEST_SAGEMAKER""" , """False""" ) ) is not True , reason="""Skipping test because should only be run when releasing minor transformers version""" , )
@pytest.mark.usefixtures("""sm_env""" )
@parameterized_class(
[
{
"""framework""": """pytorch""",
"""script""": """run_glue_model_parallelism.py""",
"""model_name_or_path""": """roberta-large""",
"""instance_type""": """ml.p3dn.24xlarge""",
"""results""": {"""train_runtime""": 1_6_0_0, """eval_accuracy""": 0.3, """eval_loss""": 1.2},
},
{
"""framework""": """pytorch""",
"""script""": """run_glue.py""",
"""model_name_or_path""": """roberta-large""",
"""instance_type""": """ml.p3dn.24xlarge""",
"""results""": {"""train_runtime""": 1_6_0_0, """eval_accuracy""": 0.3, """eval_loss""": 1.2},
},
] )
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: Any ) -> str:
if self.framework == "pytorch":
subprocess.run(
f"""cp ./examples/pytorch/text-classification/run_glue.py {self.env.test_path}/run_glue.py""".split() , encoding="""utf-8""" , check=snake_case , )
assert hasattr(self , """env""" )
def lowerCAmelCase_ ( self: int , snake_case: Dict ) -> List[Any]:
# configuration for running training on smdistributed Model Parallel
snake_case_ :Tuple = {
"""enabled""": True,
"""processes_per_host""": 8,
}
snake_case_ :List[Any] = {
"""enabled""": True,
"""parameters""": {
"""microbatches""": 4,
"""placement_strategy""": """spread""",
"""pipeline""": """interleaved""",
"""optimize""": """speed""",
"""partitions""": 4,
"""ddp""": True,
},
}
snake_case_ :Tuple = {"""smdistributed""": {"""modelparallel""": smp_options}, """mpi""": mpi_options}
snake_case_ :Any = """trainer""" if self.script == """run_glue.py""" else """smtrainer"""
# creates estimator
return HuggingFace(
entry_point=self.script , source_dir=self.env.test_path , role=self.env.role , image_uri=self.env.image_uri , base_job_name=f"""{self.env.base_job_name}-{instance_count}-smp-{name_extension}""" , instance_count=snake_case , instance_type=self.instance_type , debugger_hook_config=snake_case , hyperparameters={
**self.env.hyperparameters,
"""model_name_or_path""": self.model_name_or_path,
"""max_steps""": 500,
} , metric_definitions=self.env.metric_definitions , distribution=snake_case , py_version="""py36""" , )
def lowerCAmelCase_ ( self: Any , snake_case: Tuple ) -> List[str]:
TrainingJobAnalytics(snake_case ).export_csv(f"""{self.env.test_path}/{job_name}_metrics.csv""" )
@parameterized.expand([(1,)] )
def lowerCAmelCase_ ( self: Dict , snake_case: Dict ) -> List[Any]:
# create estimator
snake_case_ :List[Any] = self.create_estimator(snake_case )
# run training
estimator.fit()
# result dataframe
snake_case_ :Any = TrainingJobAnalytics(estimator.latest_training_job.name ).dataframe()
# extract kpis
snake_case_ :Tuple = list(result_metrics_df[result_metrics_df.metric_name == """eval_accuracy"""]["""value"""] )
snake_case_ :Dict = list(result_metrics_df[result_metrics_df.metric_name == """eval_loss"""]["""value"""] )
# get train time from SageMaker job, this includes starting, preprocessing, stopping
snake_case_ :int = (
Session().describe_training_job(estimator.latest_training_job.name ).get("""TrainingTimeInSeconds""" , 999_999 )
)
# assert kpis
assert train_runtime <= self.results["train_runtime"]
assert all(t >= self.results["""eval_accuracy"""] for t in eval_accuracy )
assert all(t <= self.results["""eval_loss"""] for t in eval_loss )
# dump tests result into json file to share in PR
with open(f"""{estimator.latest_training_job.name}.json""" , """w""" ) as outfile:
json.dump({"""train_time""": train_runtime, """eval_accuracy""": eval_accuracy, """eval_loss""": eval_loss} , snake_case )
| 66 | 1 |
"""simple docstring"""
import inspect
import os
import unittest
import torch
import accelerate
from accelerate import Accelerator
from accelerate.test_utils import execute_subprocess_async, require_multi_gpu
from accelerate.utils import patch_environment
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: int ) -> Optional[int]:
snake_case_ :str = inspect.getfile(accelerate.test_utils )
snake_case_ :List[str] = os.path.sep.join(mod_file.split(os.path.sep )[:-1] + ["""scripts""", """test_script.py"""] )
snake_case_ :Optional[Any] = os.path.sep.join(
mod_file.split(os.path.sep )[:-1] + ["""scripts""", """test_distributed_data_loop.py"""] )
snake_case_ :str = os.path.sep.join(mod_file.split(os.path.sep )[:-1] + ["""scripts""", """test_ops.py"""] )
@require_multi_gpu
def lowerCAmelCase_ ( self: Optional[Any] ) -> List[str]:
print(f"""Found {torch.cuda.device_count()} devices.""" )
snake_case_ :Any = ["""torchrun""", f"""--nproc_per_node={torch.cuda.device_count()}""", self.test_file_path]
with patch_environment(omp_num_threads=1 ):
execute_subprocess_async(snake_case , env=os.environ.copy() )
@require_multi_gpu
def lowerCAmelCase_ ( self: Optional[Any] ) -> str:
print(f"""Found {torch.cuda.device_count()} devices.""" )
snake_case_ :str = ["""torchrun""", f"""--nproc_per_node={torch.cuda.device_count()}""", self.operation_file_path]
print(f"""Command: {cmd}""" )
with patch_environment(omp_num_threads=1 ):
execute_subprocess_async(snake_case , env=os.environ.copy() )
@require_multi_gpu
def lowerCAmelCase_ ( self: Union[str, Any] ) -> List[str]:
snake_case_ :List[Any] = ["""torchrun""", f"""--nproc_per_node={torch.cuda.device_count()}""", inspect.getfile(self.__class__ )]
with patch_environment(omp_num_threads=1 ):
execute_subprocess_async(snake_case , env=os.environ.copy() )
@require_multi_gpu
def lowerCAmelCase_ ( self: str ) -> List[str]:
print(f"""Found {torch.cuda.device_count()} devices, using 2 devices only""" )
snake_case_ :Optional[int] = ["""torchrun""", f"""--nproc_per_node={torch.cuda.device_count()}""", self.data_loop_file_path]
with patch_environment(omp_num_threads=1 , cuda_visible_devices="""0,1""" ):
execute_subprocess_async(snake_case , env=os.environ.copy() )
if __name__ == "__main__":
__a = Accelerator()
__a = (accelerator.state.process_index + 2, 10)
__a = torch.randint(0, 10, shape).to(accelerator.device)
__a = ""
__a = accelerator.pad_across_processes(tensor)
if tensora.shape[0] != accelerator.state.num_processes + 1:
error_msg += F"Found shape {tensora.shape} but should have {accelerator.state.num_processes + 1} at dim 0."
if not torch.equal(tensora[: accelerator.state.process_index + 2], tensor):
error_msg += "Tensors have different values."
if not torch.all(tensora[accelerator.state.process_index + 2 :] == 0):
error_msg += "Padding was not done with the right value (0)."
__a = accelerator.pad_across_processes(tensor, pad_first=True)
if tensora.shape[0] != accelerator.state.num_processes + 1:
error_msg += F"Found shape {tensora.shape} but should have {accelerator.state.num_processes + 1} at dim 0."
__a = accelerator.state.num_processes - accelerator.state.process_index - 1
if not torch.equal(tensora[index:], tensor):
error_msg += "Tensors have different values."
if not torch.all(tensora[:index] == 0):
error_msg += "Padding was not done with the right value (0)."
# Raise error at the end to make sure we don't stop at the first failure.
if len(error_msg) > 0:
raise ValueError(error_msg)
| 66 |
"""simple docstring"""
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 lowerCamelCase :
'''simple docstring'''
def __init__( self: Dict , snake_case: Optional[Any] , snake_case: Tuple=13 , snake_case: Any=32 , snake_case: Union[str, Any]=2 , snake_case: Tuple=3 , snake_case: Union[str, Any]=16 , snake_case: Union[str, Any]=[1, 2, 1] , snake_case: Optional[Any]=[2, 2, 4] , snake_case: str=2 , snake_case: List[str]=2.0 , snake_case: Optional[int]=True , snake_case: Union[str, Any]=0.0 , snake_case: Optional[int]=0.0 , snake_case: Optional[Any]=0.1 , snake_case: List[str]="gelu" , snake_case: Any=False , snake_case: Optional[Any]=True , snake_case: Optional[int]=0.0_2 , snake_case: Any=1E-5 , snake_case: Optional[int]=True , snake_case: int=None , snake_case: Any=True , snake_case: str=10 , snake_case: Optional[Any]=8 , snake_case: Union[str, Any]=["stage1", "stage2", "stage3"] , snake_case: Tuple=[1, 2, 3] , ) -> Dict:
snake_case_ :Dict = parent
snake_case_ :List[Any] = batch_size
snake_case_ :Dict = image_size
snake_case_ :Dict = patch_size
snake_case_ :Tuple = num_channels
snake_case_ :List[Any] = embed_dim
snake_case_ :List[str] = depths
snake_case_ :str = num_heads
snake_case_ :Tuple = window_size
snake_case_ :Tuple = mlp_ratio
snake_case_ :int = qkv_bias
snake_case_ :Tuple = hidden_dropout_prob
snake_case_ :Optional[Any] = attention_probs_dropout_prob
snake_case_ :Dict = drop_path_rate
snake_case_ :Any = hidden_act
snake_case_ :Any = use_absolute_embeddings
snake_case_ :int = patch_norm
snake_case_ :List[Any] = layer_norm_eps
snake_case_ :Tuple = initializer_range
snake_case_ :str = is_training
snake_case_ :int = scope
snake_case_ :Tuple = use_labels
snake_case_ :Tuple = type_sequence_label_size
snake_case_ :str = encoder_stride
snake_case_ :List[Any] = out_features
snake_case_ :str = out_indices
def lowerCAmelCase_ ( self: Tuple ) -> Dict:
snake_case_ :Optional[int] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case_ :str = None
if self.use_labels:
snake_case_ :Optional[int] = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ :Union[str, Any] = self.get_config()
return config, pixel_values, labels
def lowerCAmelCase_ ( self: int ) -> Optional[Any]:
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 lowerCAmelCase_ ( self: List[Any] , snake_case: str , snake_case: int , snake_case: List[str] ) -> Any:
snake_case_ :Dict = MaskFormerSwinModel(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Tuple = model(snake_case )
snake_case_ :Dict = ((config.image_size // config.patch_size) ** 2) // (4 ** (len(config.depths ) - 1))
snake_case_ :Any = 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 lowerCAmelCase_ ( self: Optional[Any] , snake_case: int , snake_case: List[str] , snake_case: Tuple ) -> Union[str, Any]:
snake_case_ :Any = MaskFormerSwinBackbone(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Optional[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_ :Optional[Any] = ["""stem"""]
snake_case_ :str = MaskFormerSwinBackbone(config=snake_case )
def lowerCAmelCase_ ( self: List[str] ) -> Optional[Any]:
snake_case_ :Optional[int] = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_ :str = config_and_inputs
snake_case_ :Tuple = {"""pixel_values""": pixel_values}
return config, inputs_dict
@require_torch
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : Union[str, Any] = (
(
MaskFormerSwinModel,
MaskFormerSwinBackbone,
)
if is_torch_available()
else ()
)
_A : str = {"""feature-extraction""": MaskFormerSwinModel} if is_torch_available() else {}
_A : List[str] = False
_A : Any = False
_A : Dict = False
_A : List[Any] = False
_A : Optional[int] = False
def lowerCAmelCase_ ( self: Dict ) -> Any:
snake_case_ :str = MaskFormerSwinModelTester(self )
snake_case_ :Optional[Any] = 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 lowerCAmelCase_ ( self: List[str] ) -> Optional[int]:
pass
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Dict:
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 lowerCAmelCase_ ( self: Any ) -> Tuple:
return
def lowerCAmelCase_ ( self: Any ) -> Any:
snake_case_ :List[Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> int:
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 lowerCAmelCase_ ( self: str ) -> List[str]:
pass
@unittest.skip("""Swin does not support feedforward chunking""" )
def lowerCAmelCase_ ( self: int ) -> Optional[int]:
pass
def lowerCAmelCase_ ( self: List[str] ) -> List[Any]:
snake_case_, snake_case_ :List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :str = model_class(snake_case )
self.assertIsInstance(model.get_input_embeddings() , (nn.Module) )
snake_case_ :Dict = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(snake_case , nn.Linear ) )
def lowerCAmelCase_ ( self: Tuple ) -> Dict:
snake_case_, snake_case_ :int = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :Optional[int] = model_class(snake_case )
snake_case_ :str = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case_ :str = [*signature.parameters.keys()]
snake_case_ :str = ["""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 lowerCAmelCase_ ( self: List[Any] ) -> List[Any]:
pass
@unittest.skip(reason="""MaskFormerSwin is only used as an internal backbone""" )
def lowerCAmelCase_ ( self: Dict ) -> List[Any]:
pass
def lowerCAmelCase_ ( self: Union[str, Any] , snake_case: Union[str, Any] , snake_case: int , snake_case: Any , snake_case: List[str] ) -> str:
snake_case_ :List[str] = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :List[Any] = model(**self._prepare_for_class(snake_case , snake_case ) )
snake_case_ :Any = outputs.hidden_states
snake_case_ :Optional[int] = 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_ :str = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :int = (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 lowerCAmelCase_ ( self: List[Any] ) -> Optional[int]:
snake_case_, snake_case_ :Any = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :List[Any] = (
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_ :Tuple = 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_ :List[Any] = True
self.check_hidden_states_output(snake_case , snake_case , snake_case , snake_case )
def lowerCAmelCase_ ( self: Optional[Any] ) -> Tuple:
snake_case_, snake_case_ :int = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :List[Any] = 3
snake_case_ :List[Any] = (
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_ :Any = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :Tuple = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
snake_case_ :List[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_ :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 lowerCAmelCase_ ( self: Union[str, Any] ) -> List[str]:
pass
@unittest.skip(reason="""This will be fixed once MaskFormerSwin is replaced by native Swin""" )
def lowerCAmelCase_ ( self: List[str] ) -> str:
pass
@unittest.skip(reason="""This will be fixed once MaskFormerSwin is replaced by native Swin""" )
def lowerCAmelCase_ ( self: str ) -> List[Any]:
pass
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Optional[Any]:
snake_case_, snake_case_ :Dict = self.model_tester.prepare_config_and_inputs_for_common()
def set_nan_tensor_to_zero(snake_case: str ):
snake_case_ :Optional[int] = 0
return t
def check_equivalence(snake_case: List[Any] , snake_case: Union[str, Any] , snake_case: int , snake_case: Tuple={} ):
with torch.no_grad():
snake_case_ :List[Any] = model(**snake_case , return_dict=snake_case , **snake_case )
snake_case_ :Any = model(**snake_case , return_dict=snake_case , **snake_case ).to_tuple()
def recursive_check(snake_case: List[Any] , snake_case: int ):
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_ :int = model_class(snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Any = 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 )
snake_case_ :Tuple = self._prepare_for_class(snake_case , snake_case , return_labels=snake_case )
snake_case_ :Dict = self._prepare_for_class(snake_case , snake_case , return_labels=snake_case )
check_equivalence(snake_case , snake_case , snake_case )
snake_case_ :Tuple = self._prepare_for_class(snake_case , snake_case )
snake_case_ :Any = self._prepare_for_class(snake_case , snake_case )
check_equivalence(snake_case , snake_case , snake_case , {"""output_hidden_states""": True} )
snake_case_ :Dict = 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 lowerCamelCase ( unittest.TestCase , _lowerCAmelCase ):
'''simple docstring'''
_A : int = (MaskFormerSwinBackbone,) if is_torch_available() else ()
_A : Tuple = MaskFormerSwinConfig
def lowerCAmelCase_ ( self: List[str] ) -> Optional[int]:
snake_case_ :Optional[Any] = MaskFormerSwinModelTester(self )
def lowerCAmelCase_ ( self: int ) -> Optional[int]:
snake_case_, snake_case_ :Any = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Tuple = inputs_dict["""pixel_values"""].shape[0]
for backbone_class in self.all_model_classes:
snake_case_ :List[str] = backbone_class(snake_case )
backbone.to(snake_case )
backbone.eval()
snake_case_ :List[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_ :Union[str, Any] = 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_ :List[Any] = 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_ :List[Any] = backbone(**snake_case , output_attentions=snake_case )
self.assertIsNotNone(outputs.attentions )
| 66 | 1 |
"""simple docstring"""
import json
import os
import unittest
from transformers import MgpstrTokenizer
from transformers.models.mgp_str.tokenization_mgp_str import VOCAB_FILES_NAMES
from transformers.testing_utils import require_tokenizers
from ...test_tokenization_common import TokenizerTesterMixin
@require_tokenizers
class lowerCamelCase ( _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : Optional[int] = MgpstrTokenizer
_A : Tuple = False
_A : str = {}
_A : Any = False
def lowerCAmelCase_ ( self: List[Any] ) -> List[str]:
super().setUp()
# fmt: off
snake_case_ :int = ["""[GO]""", """[s]""", """0""", """1""", """2""", """3""", """4""", """5""", """6""", """7""", """8""", """9""", """a""", """b""", """c""", """d""", """e""", """f""", """g""", """h""", """i""", """j""", """k""", """l""", """m""", """n""", """o""", """p""", """q""", """r""", """s""", """t""", """u""", """v""", """w""", """x""", """y""", """z"""]
# fmt: on
snake_case_ :Optional[int] = dict(zip(snake_case , range(len(snake_case ) ) ) )
snake_case_ :List[str] = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES["""vocab_file"""] )
with open(self.vocab_file , """w""" , encoding="""utf-8""" ) as fp:
fp.write(json.dumps(snake_case ) + """\n""" )
def lowerCAmelCase_ ( self: Optional[int] , **snake_case: Union[str, Any] ) -> Optional[Any]:
return MgpstrTokenizer.from_pretrained(self.tmpdirname , **snake_case )
def lowerCAmelCase_ ( self: int , snake_case: Any ) -> Optional[int]:
snake_case_ :Optional[int] = """tester"""
snake_case_ :Union[str, Any] = """tester"""
return input_text, output_text
@unittest.skip("""MGP-STR always lower cases letters.""" )
def lowerCAmelCase_ ( self: List[Any] ) -> List[str]:
pass
def lowerCAmelCase_ ( self: Optional[int] ) -> int:
snake_case_ :List[Any] = self.get_tokenizers(do_lower_case=snake_case )
for tokenizer in tokenizers:
with self.subTest(f"""{tokenizer.__class__.__name__}""" ):
snake_case_ :List[str] = """[SPECIAL_TOKEN]"""
tokenizer.add_special_tokens({"""cls_token""": special_token} )
snake_case_ :Any = tokenizer.encode([special_token] , add_special_tokens=snake_case )
self.assertEqual(len(snake_case ) , 1 )
snake_case_ :str = tokenizer.decode(snake_case , skip_special_tokens=snake_case )
self.assertTrue(special_token not in decoded )
def lowerCAmelCase_ ( self: Optional[Any] ) -> int:
snake_case_ :Dict = self.get_tokenizers()
for tokenizer in tokenizers:
with self.subTest(f"""{tokenizer.__class__.__name__}""" ):
snake_case_, snake_case_ :Any = self.get_input_output_texts(snake_case )
snake_case_ :Any = tokenizer.tokenize(snake_case )
snake_case_ :Optional[int] = tokenizer.convert_tokens_to_ids(snake_case )
snake_case_ :Optional[int] = tokenizer.encode(snake_case , add_special_tokens=snake_case )
self.assertListEqual(snake_case , snake_case )
snake_case_ :int = tokenizer.convert_ids_to_tokens(snake_case )
self.assertNotEqual(len(snake_case ) , 0 )
snake_case_ :Optional[int] = tokenizer.decode(snake_case )
self.assertIsInstance(snake_case , snake_case )
self.assertEqual(text_a.replace(""" """ , """""" ) , snake_case )
@unittest.skip("""MGP-STR tokenizer only handles one sequence.""" )
def lowerCAmelCase_ ( self: Any ) -> Optional[int]:
pass
@unittest.skip("""inputs cannot be pretokenized in MgpstrTokenizer""" )
def lowerCAmelCase_ ( self: Optional[Any] ) -> Optional[int]:
pass
| 66 |
"""simple docstring"""
import gc
import math
import unittest
import torch
from diffusers import UNetaDModel
from diffusers.utils import floats_tensor, logging, slow, torch_all_close, torch_device
from diffusers.utils.testing_utils import enable_full_determinism
from .test_modeling_common import ModelTesterMixin, UNetTesterMixin
__a = logging.get_logger(__name__)
enable_full_determinism()
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[Any] = UNetaDModel
_A : Union[str, Any] = """sample"""
@property
def lowerCAmelCase_ ( self: str ) -> Tuple:
snake_case_ :List[str] = 4
snake_case_ :Tuple = 3
snake_case_ :Optional[Any] = (32, 32)
snake_case_ :str = floats_tensor((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :Union[str, Any] = torch.tensor([10] ).to(snake_case )
return {"sample": noise, "timestep": time_step}
@property
def lowerCAmelCase_ ( self: List[str] ) -> Dict:
return (3, 32, 32)
@property
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
return (3, 32, 32)
def lowerCAmelCase_ ( self: Optional[int] ) -> Dict:
snake_case_ :Any = {
"""block_out_channels""": (32, 64),
"""down_block_types""": ("""DownBlock2D""", """AttnDownBlock2D"""),
"""up_block_types""": ("""AttnUpBlock2D""", """UpBlock2D"""),
"""attention_head_dim""": 3,
"""out_channels""": 3,
"""in_channels""": 3,
"""layers_per_block""": 2,
"""sample_size""": 32,
}
snake_case_ :Tuple = self.dummy_input
return init_dict, inputs_dict
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[str] = UNetaDModel
_A : Union[str, Any] = """sample"""
@property
def lowerCAmelCase_ ( self: str ) -> str:
snake_case_ :List[str] = 4
snake_case_ :Optional[int] = 4
snake_case_ :int = (32, 32)
snake_case_ :Any = floats_tensor((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :List[Any] = torch.tensor([10] ).to(snake_case )
return {"sample": noise, "timestep": time_step}
@property
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
return (4, 32, 32)
@property
def lowerCAmelCase_ ( self: List[Any] ) -> int:
return (4, 32, 32)
def lowerCAmelCase_ ( self: Union[str, Any] ) -> List[Any]:
snake_case_ :Dict = {
"""sample_size""": 32,
"""in_channels""": 4,
"""out_channels""": 4,
"""layers_per_block""": 2,
"""block_out_channels""": (32, 64),
"""attention_head_dim""": 32,
"""down_block_types""": ("""DownBlock2D""", """DownBlock2D"""),
"""up_block_types""": ("""UpBlock2D""", """UpBlock2D"""),
}
snake_case_ :List[str] = self.dummy_input
return init_dict, inputs_dict
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[Any]:
snake_case_, snake_case_ :List[str] = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case )
self.assertIsNotNone(snake_case )
self.assertEqual(len(loading_info["""missing_keys"""] ) , 0 )
model.to(snake_case )
snake_case_ :List[str] = model(**self.dummy_input ).sample
assert image is not None, "Make sure output is not None"
@unittest.skipIf(torch_device != """cuda""" , """This test is supposed to run on GPU""" )
def lowerCAmelCase_ ( self: Tuple ) -> Dict:
snake_case_, snake_case_ :Union[str, Any] = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case )
model.to(snake_case )
snake_case_ :Union[str, Any] = model(**self.dummy_input ).sample
assert image is not None, "Make sure output is not None"
@unittest.skipIf(torch_device != """cuda""" , """This test is supposed to run on GPU""" )
def lowerCAmelCase_ ( self: str ) -> Any:
# by defautl model loading will use accelerate as `low_cpu_mem_usage=True`
snake_case_, snake_case_ :List[str] = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case )
model_accelerate.to(snake_case )
model_accelerate.eval()
snake_case_ :List[Any] = torch.randn(
1 , model_accelerate.config.in_channels , model_accelerate.config.sample_size , model_accelerate.config.sample_size , generator=torch.manual_seed(0 ) , )
snake_case_ :int = noise.to(snake_case )
snake_case_ :str = torch.tensor([10] * noise.shape[0] ).to(snake_case )
snake_case_ :Optional[int] = model_accelerate(snake_case , snake_case )["""sample"""]
# two models don't need to stay in the device at the same time
del model_accelerate
torch.cuda.empty_cache()
gc.collect()
snake_case_, snake_case_ :str = UNetaDModel.from_pretrained(
"""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case , low_cpu_mem_usage=snake_case )
model_normal_load.to(snake_case )
model_normal_load.eval()
snake_case_ :int = model_normal_load(snake_case , snake_case )["""sample"""]
assert torch_all_close(snake_case , snake_case , rtol=1E-3 )
def lowerCAmelCase_ ( self: Tuple ) -> Any:
snake_case_ :Tuple = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" )
model.eval()
model.to(snake_case )
snake_case_ :Optional[int] = torch.randn(
1 , model.config.in_channels , model.config.sample_size , model.config.sample_size , generator=torch.manual_seed(0 ) , )
snake_case_ :int = noise.to(snake_case )
snake_case_ :List[Any] = torch.tensor([10] * noise.shape[0] ).to(snake_case )
with torch.no_grad():
snake_case_ :Union[str, Any] = model(snake_case , snake_case ).sample
snake_case_ :Optional[int] = output[0, -1, -3:, -3:].flatten().cpu()
# fmt: off
snake_case_ :Dict = torch.tensor([-1_3.3_2_5_8, -2_0.1_1_0_0, -1_5.9_8_7_3, -1_7.6_6_1_7, -2_3.0_5_9_6, -1_7.9_4_1_9, -1_3.3_6_7_5, -1_6.1_8_8_9, -1_2.3_8_0_0] )
# fmt: on
self.assertTrue(torch_all_close(snake_case , snake_case , rtol=1E-3 ) )
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[Any] = UNetaDModel
_A : List[Any] = """sample"""
@property
def lowerCAmelCase_ ( self: Union[str, Any] , snake_case: int=(32, 32) ) -> Tuple:
snake_case_ :Union[str, Any] = 4
snake_case_ :Any = 3
snake_case_ :int = floats_tensor((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :Any = torch.tensor(batch_size * [10] ).to(dtype=torch.intaa , device=snake_case )
return {"sample": noise, "timestep": time_step}
@property
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Any:
return (3, 32, 32)
@property
def lowerCAmelCase_ ( self: int ) -> Tuple:
return (3, 32, 32)
def lowerCAmelCase_ ( self: List[str] ) -> Tuple:
snake_case_ :List[Any] = {
"""block_out_channels""": [32, 64, 64, 64],
"""in_channels""": 3,
"""layers_per_block""": 1,
"""out_channels""": 3,
"""time_embedding_type""": """fourier""",
"""norm_eps""": 1E-6,
"""mid_block_scale_factor""": math.sqrt(2.0 ),
"""norm_num_groups""": None,
"""down_block_types""": [
"""SkipDownBlock2D""",
"""AttnSkipDownBlock2D""",
"""SkipDownBlock2D""",
"""SkipDownBlock2D""",
],
"""up_block_types""": [
"""SkipUpBlock2D""",
"""SkipUpBlock2D""",
"""AttnSkipUpBlock2D""",
"""SkipUpBlock2D""",
],
}
snake_case_ :int = self.dummy_input
return init_dict, inputs_dict
@slow
def lowerCAmelCase_ ( self: Optional[Any] ) -> List[Any]:
snake_case_, snake_case_ :List[Any] = UNetaDModel.from_pretrained("""google/ncsnpp-celebahq-256""" , output_loading_info=snake_case )
self.assertIsNotNone(snake_case )
self.assertEqual(len(loading_info["""missing_keys"""] ) , 0 )
model.to(snake_case )
snake_case_ :Any = self.dummy_input
snake_case_ :int = floats_tensor((4, 3) + (256, 256) ).to(snake_case )
snake_case_ :int = noise
snake_case_ :int = model(**snake_case )
assert image is not None, "Make sure output is not None"
@slow
def lowerCAmelCase_ ( self: str ) -> Dict:
snake_case_ :Dict = UNetaDModel.from_pretrained("""google/ncsnpp-celebahq-256""" )
model.to(snake_case )
snake_case_ :List[str] = 4
snake_case_ :Optional[int] = 3
snake_case_ :List[str] = (256, 256)
snake_case_ :Tuple = torch.ones((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :str = torch.tensor(batch_size * [1E-4] ).to(snake_case )
with torch.no_grad():
snake_case_ :Dict = model(snake_case , snake_case ).sample
snake_case_ :int = output[0, -3:, -3:, -1].flatten().cpu()
# fmt: off
snake_case_ :Optional[Any] = torch.tensor([-4_8_4_2.8_6_9_1, -6_4_9_9.6_6_3_1, -3_8_0_0.1_9_5_3, -7_9_7_8.2_6_8_6, -1_0_9_8_0.7_1_2_9, -2_0_0_2_8.8_5_3_5, 8_1_4_8.2_8_2_2, 2_3_4_2.2_9_0_5, 5_6_7.7_6_0_8] )
# fmt: on
self.assertTrue(torch_all_close(snake_case , snake_case , rtol=1E-2 ) )
def lowerCAmelCase_ ( self: List[str] ) -> List[Any]:
snake_case_ :Optional[Any] = UNetaDModel.from_pretrained("""fusing/ncsnpp-ffhq-ve-dummy-update""" )
model.to(snake_case )
snake_case_ :Optional[int] = 4
snake_case_ :Optional[Any] = 3
snake_case_ :Optional[Any] = (32, 32)
snake_case_ :Dict = torch.ones((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :Any = torch.tensor(batch_size * [1E-4] ).to(snake_case )
with torch.no_grad():
snake_case_ :str = model(snake_case , snake_case ).sample
snake_case_ :int = output[0, -3:, -3:, -1].flatten().cpu()
# fmt: off
snake_case_ :int = torch.tensor([-0.0_3_2_5, -0.0_9_0_0, -0.0_8_6_9, -0.0_3_3_2, -0.0_7_2_5, -0.0_2_7_0, -0.0_1_0_1, 0.0_2_2_7, 0.0_2_5_6] )
# fmt: on
self.assertTrue(torch_all_close(snake_case , snake_case , rtol=1E-2 ) )
def lowerCAmelCase_ ( self: Dict ) -> Optional[Any]:
# not required for this model
pass
| 66 | 1 |
"""simple docstring"""
import gc
import unittest
from diffusers import FlaxControlNetModel, FlaxStableDiffusionControlNetPipeline
from diffusers.utils import is_flax_available, load_image, slow
from diffusers.utils.testing_utils import require_flax
if is_flax_available():
import jax
import jax.numpy as jnp
from flax.jax_utils import replicate
from flax.training.common_utils import shard
@slow
@require_flax
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: List[Any] ) -> Any:
# clean up the VRAM after each test
super().tearDown()
gc.collect()
def lowerCAmelCase_ ( self: Tuple ) -> Any:
snake_case_, snake_case_ :List[str] = FlaxControlNetModel.from_pretrained(
"""lllyasviel/sd-controlnet-canny""" , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_, snake_case_ :Union[str, Any] = FlaxStableDiffusionControlNetPipeline.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" , controlnet=snake_case , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_ :Union[str, Any] = controlnet_params
snake_case_ :Union[str, Any] = """bird"""
snake_case_ :List[Any] = jax.device_count()
snake_case_ :List[Any] = pipe.prepare_text_inputs([prompts] * num_samples )
snake_case_ :List[str] = load_image(
"""https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/sd_controlnet/bird_canny.png""" )
snake_case_ :List[str] = pipe.prepare_image_inputs([canny_image] * num_samples )
snake_case_ :Any = jax.random.PRNGKey(0 )
snake_case_ :List[str] = jax.random.split(snake_case , jax.device_count() )
snake_case_ :List[Any] = replicate(snake_case )
snake_case_ :List[str] = shard(snake_case )
snake_case_ :str = shard(snake_case )
snake_case_ :Dict = pipe(
prompt_ids=snake_case , image=snake_case , params=snake_case , prng_seed=snake_case , num_inference_steps=50 , jit=snake_case , ).images
assert images.shape == (jax.device_count(), 1, 768, 512, 3)
snake_case_ :str = images.reshape((images.shape[0] * images.shape[1],) + images.shape[-3:] )
snake_case_ :Union[str, Any] = images[0, 253:256, 253:256, -1]
snake_case_ :str = jnp.asarray(jax.device_get(image_slice.flatten() ) )
snake_case_ :Dict = jnp.array(
[0.1_6_7_9_6_9, 0.1_1_6_6_9_9, 0.0_8_1_5_4_3, 0.1_5_4_2_9_7, 0.1_3_2_8_1_2, 0.1_0_8_8_8_7, 0.1_6_9_9_2_2, 0.1_6_9_9_2_2, 0.2_0_5_0_7_8] )
print(f"""output_slice: {output_slice}""" )
assert jnp.abs(output_slice - expected_slice ).max() < 1E-2
def lowerCAmelCase_ ( self: int ) -> Dict:
snake_case_, snake_case_ :List[Any] = FlaxControlNetModel.from_pretrained(
"""lllyasviel/sd-controlnet-openpose""" , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_, snake_case_ :int = FlaxStableDiffusionControlNetPipeline.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" , controlnet=snake_case , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_ :str = controlnet_params
snake_case_ :Optional[int] = """Chef in the kitchen"""
snake_case_ :Union[str, Any] = jax.device_count()
snake_case_ :Any = pipe.prepare_text_inputs([prompts] * num_samples )
snake_case_ :str = load_image(
"""https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/sd_controlnet/pose.png""" )
snake_case_ :Optional[Any] = pipe.prepare_image_inputs([pose_image] * num_samples )
snake_case_ :str = jax.random.PRNGKey(0 )
snake_case_ :str = jax.random.split(snake_case , jax.device_count() )
snake_case_ :Tuple = replicate(snake_case )
snake_case_ :str = shard(snake_case )
snake_case_ :int = shard(snake_case )
snake_case_ :List[str] = pipe(
prompt_ids=snake_case , image=snake_case , params=snake_case , prng_seed=snake_case , num_inference_steps=50 , jit=snake_case , ).images
assert images.shape == (jax.device_count(), 1, 768, 512, 3)
snake_case_ :str = images.reshape((images.shape[0] * images.shape[1],) + images.shape[-3:] )
snake_case_ :int = images[0, 253:256, 253:256, -1]
snake_case_ :Dict = jnp.asarray(jax.device_get(image_slice.flatten() ) )
snake_case_ :Optional[int] = jnp.array(
[[0.2_7_1_4_8_4, 0.2_6_1_7_1_9, 0.2_7_5_3_9_1, 0.2_7_7_3_4_4, 0.2_7_9_2_9_7, 0.2_9_1_0_1_6, 0.2_9_4_9_2_2, 0.3_0_2_7_3_4, 0.3_0_2_7_3_4]] )
print(f"""output_slice: {output_slice}""" )
assert jnp.abs(output_slice - expected_slice ).max() < 1E-2
| 66 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available, is_vision_available
__a = {
"configuration_mask2former": [
"MASK2FORMER_PRETRAINED_CONFIG_ARCHIVE_MAP",
"Mask2FormerConfig",
],
}
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["Mask2FormerImageProcessor"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"MASK2FORMER_PRETRAINED_MODEL_ARCHIVE_LIST",
"Mask2FormerForUniversalSegmentation",
"Mask2FormerModel",
"Mask2FormerPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_maskaformer import MASK2FORMER_PRETRAINED_CONFIG_ARCHIVE_MAP, MaskaFormerConfig
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .image_processing_maskaformer import MaskaFormerImageProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_maskaformer import (
MASK2FORMER_PRETRAINED_MODEL_ARCHIVE_LIST,
MaskaFormerForUniversalSegmentation,
MaskaFormerModel,
MaskaFormerPreTrainedModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure)
| 66 | 1 |
"""simple docstring"""
import os
import unittest
from transformers.models.transfo_xl.tokenization_transfo_xl import VOCAB_FILES_NAMES, TransfoXLTokenizer
from ...test_tokenization_common import TokenizerTesterMixin
class lowerCamelCase ( _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : Optional[int] = TransfoXLTokenizer
_A : int = False
_A : Any = False
def lowerCAmelCase_ ( self: int ) -> Tuple:
super().setUp()
snake_case_ :Optional[int] = [
"""<unk>""",
"""[CLS]""",
"""[SEP]""",
"""want""",
"""unwanted""",
"""wa""",
"""un""",
"""running""",
""",""",
"""low""",
"""l""",
]
snake_case_ :List[Any] = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES["""vocab_file"""] )
with open(self.vocab_file , """w""" , encoding="""utf-8""" ) as vocab_writer:
vocab_writer.write("""""".join([x + """\n""" for x in vocab_tokens] ) )
def lowerCAmelCase_ ( self: Optional[Any] , **snake_case: Optional[int] ) -> Dict:
snake_case_ :Union[str, Any] = True
return TransfoXLTokenizer.from_pretrained(self.tmpdirname , **snake_case )
def lowerCAmelCase_ ( self: str , snake_case: List[str] ) -> Tuple:
snake_case_ :List[Any] = """<unk> UNwanted , running"""
snake_case_ :Union[str, Any] = """<unk> unwanted, running"""
return input_text, output_text
def lowerCAmelCase_ ( self: int ) -> str:
snake_case_ :str = TransfoXLTokenizer(vocab_file=self.vocab_file , lower_case=snake_case )
snake_case_ :str = tokenizer.tokenize("""<unk> UNwanted , running""" )
self.assertListEqual(snake_case , ["""<unk>""", """unwanted""", """,""", """running"""] )
self.assertListEqual(tokenizer.convert_tokens_to_ids(snake_case ) , [0, 4, 8, 7] )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Tuple:
snake_case_ :Union[str, Any] = TransfoXLTokenizer(lower_case=snake_case )
self.assertListEqual(
tokenizer.tokenize(""" \tHeLLo ! how \n Are yoU ? """ ) , ["""hello""", """!""", """how""", """are""", """you""", """?"""] )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Any:
snake_case_ :List[str] = TransfoXLTokenizer(lower_case=snake_case )
self.assertListEqual(
tokenizer.tokenize(""" \tHeLLo ! how \n Are yoU ? """ ) , ["""HeLLo""", """!""", """how""", """Are""", """yoU""", """?"""] )
def lowerCAmelCase_ ( self: List[Any] ) -> Optional[int]:
snake_case_ :Tuple = TransfoXLTokenizer(lower_case=snake_case )
snake_case_ :int = """Hello (bracket) and side-scrolled [and] Henry's $5,000 with 3.34 m. What's up!?"""
snake_case_ :str = [
"""Hello""",
"""(""",
"""bracket""",
""")""",
"""and""",
"""side""",
"""@-@""",
"""scrolled""",
"""[""",
"""and""",
"""]""",
"""Henry""",
"""'s""",
"""$""",
"""5""",
"""@,@""",
"""000""",
"""with""",
"""3""",
"""@.@""",
"""34""",
"""m""",
""".""",
"""What""",
"""'s""",
"""up""",
"""!""",
"""?""",
]
self.assertListEqual(tokenizer.tokenize(snake_case ) , snake_case )
self.assertEqual(tokenizer.convert_tokens_to_string(snake_case ) , snake_case )
def lowerCAmelCase_ ( self: Optional[int] ) -> Tuple:
snake_case_ :str = self.get_tokenizer()
snake_case_ :Optional[int] = len(snake_case )
tokenizer.add_tokens(["""new1""", """new2"""] )
tokenizer.move_added_token("""new1""" , 1 )
# Check that moved token is not copied (duplicate)
self.assertEqual(len(snake_case ) , original_len + 2 )
# Check that token is moved to specified id
self.assertEqual(tokenizer.encode("""new1""" ) , [1] )
self.assertEqual(tokenizer.decode([1] ) , """new1""" )
| 66 |
"""simple docstring"""
import gc
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import (
AutoencoderKL,
DDIMScheduler,
StableDiffusionSAGPipeline,
UNetaDConditionModel,
)
from diffusers.utils import slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu
from ..pipeline_params import TEXT_TO_IMAGE_BATCH_PARAMS, TEXT_TO_IMAGE_IMAGE_PARAMS, TEXT_TO_IMAGE_PARAMS
from ..test_pipelines_common import PipelineLatentTesterMixin, PipelineTesterMixin
enable_full_determinism()
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : str = StableDiffusionSAGPipeline
_A : Optional[Any] = TEXT_TO_IMAGE_PARAMS
_A : Any = TEXT_TO_IMAGE_BATCH_PARAMS
_A : Tuple = TEXT_TO_IMAGE_IMAGE_PARAMS
_A : Tuple = TEXT_TO_IMAGE_IMAGE_PARAMS
_A : List[str] = False
def lowerCAmelCase_ ( self: Optional[Any] ) -> str:
torch.manual_seed(0 )
snake_case_ :Any = UNetaDConditionModel(
block_out_channels=(32, 64) , layers_per_block=2 , sample_size=32 , in_channels=4 , out_channels=4 , down_block_types=("""DownBlock2D""", """CrossAttnDownBlock2D""") , up_block_types=("""CrossAttnUpBlock2D""", """UpBlock2D""") , cross_attention_dim=32 , )
snake_case_ :Any = DDIMScheduler(
beta_start=0.0_0_0_8_5 , beta_end=0.0_1_2 , beta_schedule="""scaled_linear""" , clip_sample=snake_case , set_alpha_to_one=snake_case , )
torch.manual_seed(0 )
snake_case_ :Optional[int] = AutoencoderKL(
block_out_channels=[32, 64] , in_channels=3 , out_channels=3 , down_block_types=["""DownEncoderBlock2D""", """DownEncoderBlock2D"""] , up_block_types=["""UpDecoderBlock2D""", """UpDecoderBlock2D"""] , latent_channels=4 , )
torch.manual_seed(0 )
snake_case_ :Union[str, Any] = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1E-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1_000 , )
snake_case_ :Tuple = CLIPTextModel(snake_case )
snake_case_ :str = CLIPTokenizer.from_pretrained("""hf-internal-testing/tiny-random-clip""" )
snake_case_ :Dict = {
"""unet""": unet,
"""scheduler""": scheduler,
"""vae""": vae,
"""text_encoder""": text_encoder,
"""tokenizer""": tokenizer,
"""safety_checker""": None,
"""feature_extractor""": None,
}
return components
def lowerCAmelCase_ ( self: List[str] , snake_case: Tuple , snake_case: List[str]=0 ) -> str:
if str(snake_case ).startswith("""mps""" ):
snake_case_ :Tuple = torch.manual_seed(snake_case )
else:
snake_case_ :Optional[int] = torch.Generator(device=snake_case ).manual_seed(snake_case )
snake_case_ :Any = {
"""prompt""": """.""",
"""generator""": generator,
"""num_inference_steps""": 2,
"""guidance_scale""": 1.0,
"""sag_scale""": 1.0,
"""output_type""": """numpy""",
}
return inputs
def lowerCAmelCase_ ( self: Optional[int] ) -> str:
super().test_inference_batch_single_identical(expected_max_diff=3E-3 )
@slow
@require_torch_gpu
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: int ) -> str:
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def lowerCAmelCase_ ( self: int ) -> List[str]:
snake_case_ :Any = StableDiffusionSAGPipeline.from_pretrained("""CompVis/stable-diffusion-v1-4""" )
snake_case_ :int = sag_pipe.to(snake_case )
sag_pipe.set_progress_bar_config(disable=snake_case )
snake_case_ :Union[str, Any] = """."""
snake_case_ :str = torch.manual_seed(0 )
snake_case_ :str = sag_pipe(
[prompt] , generator=snake_case , guidance_scale=7.5 , sag_scale=1.0 , num_inference_steps=20 , output_type="""np""" )
snake_case_ :List[Any] = output.images
snake_case_ :Tuple = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
snake_case_ :List[Any] = np.array([0.1_5_6_8, 0.1_7_3_8, 0.1_6_9_5, 0.1_6_9_3, 0.1_5_0_7, 0.1_7_0_5, 0.1_5_4_7, 0.1_7_5_1, 0.1_9_4_9] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 5E-2
def lowerCAmelCase_ ( self: Dict ) -> str:
snake_case_ :Tuple = StableDiffusionSAGPipeline.from_pretrained("""stabilityai/stable-diffusion-2-1-base""" )
snake_case_ :Optional[int] = sag_pipe.to(snake_case )
sag_pipe.set_progress_bar_config(disable=snake_case )
snake_case_ :Tuple = """."""
snake_case_ :Union[str, Any] = torch.manual_seed(0 )
snake_case_ :Tuple = sag_pipe(
[prompt] , generator=snake_case , guidance_scale=7.5 , sag_scale=1.0 , num_inference_steps=20 , output_type="""np""" )
snake_case_ :Optional[int] = output.images
snake_case_ :Optional[Any] = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
snake_case_ :Tuple = np.array([0.3_4_5_9, 0.2_8_7_6, 0.2_5_3_7, 0.3_0_0_2, 0.2_6_7_1, 0.2_1_6_0, 0.3_0_2_6, 0.2_2_6_2, 0.2_3_7_1] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 5E-2
def lowerCAmelCase_ ( self: List[str] ) -> List[str]:
snake_case_ :Optional[int] = StableDiffusionSAGPipeline.from_pretrained("""stabilityai/stable-diffusion-2-1-base""" )
snake_case_ :int = sag_pipe.to(snake_case )
sag_pipe.set_progress_bar_config(disable=snake_case )
snake_case_ :Tuple = """."""
snake_case_ :Optional[int] = torch.manual_seed(0 )
snake_case_ :List[str] = sag_pipe(
[prompt] , width=768 , height=512 , generator=snake_case , guidance_scale=7.5 , sag_scale=1.0 , num_inference_steps=20 , output_type="""np""" , )
snake_case_ :Optional[Any] = output.images
assert image.shape == (1, 512, 768, 3)
| 66 | 1 |
"""simple docstring"""
import argparse
import gdown
import numpy as np
import torch
from huggingface_hub import hf_hub_download
from transformers import (
CLIPTokenizer,
CLIPTokenizerFast,
VideoMAEImageProcessor,
XCLIPConfig,
XCLIPModel,
XCLIPProcessor,
XCLIPTextConfig,
XCLIPVisionConfig,
)
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :int = XCLIPTextConfig()
# derive patch size from model name
snake_case_ :Union[str, Any] = model_name.find("""patch""" )
snake_case_ :List[str] = int(model_name[start_idx + len("""patch""" ) : start_idx + len("""patch""" ) + 2] )
snake_case_ :Any = XCLIPVisionConfig(patch_size=_lowercase, num_frames=_lowercase )
if "large" in model_name:
snake_case_ :Optional[Any] = 768
snake_case_ :Union[str, Any] = 3072
snake_case_ :Any = 12
snake_case_ :Any = 1024
snake_case_ :str = 4096
snake_case_ :Union[str, Any] = 16
snake_case_ :Union[str, Any] = 24
snake_case_ :Tuple = 768
snake_case_ :Any = 3072
if model_name == "xclip-large-patch14-16-frames":
snake_case_ :Any = 336
snake_case_ :Any = XCLIPConfig.from_text_vision_configs(_lowercase, _lowercase )
if "large" in model_name:
snake_case_ :List[Any] = 768
return config
def A_ ( _lowercase ):
'''simple docstring'''
if name == "token_embedding.weight":
snake_case_ :Optional[Any] = name.replace("""token_embedding.weight""", """text_model.embeddings.token_embedding.weight""" )
if name == "positional_embedding":
snake_case_ :Tuple = name.replace("""positional_embedding""", """text_model.embeddings.position_embedding.weight""" )
if "ln_1" in name:
snake_case_ :Dict = name.replace("""ln_1""", """layer_norm1""" )
if "ln_2" in name:
snake_case_ :str = name.replace("""ln_2""", """layer_norm2""" )
if "c_fc" in name:
snake_case_ :str = name.replace("""c_fc""", """fc1""" )
if "c_proj" in name:
snake_case_ :int = name.replace("""c_proj""", """fc2""" )
if name.startswith("""transformer.resblocks""" ):
snake_case_ :Union[str, Any] = name.replace("""transformer.resblocks""", """text_model.encoder.layers""" )
if "attn.out_proj" in name and "message" not in name:
snake_case_ :Union[str, Any] = name.replace("""attn.out_proj""", """self_attn.out_proj""" )
if "ln_final" in name:
snake_case_ :Union[str, Any] = name.replace("""ln_final""", """text_model.final_layer_norm""" )
# visual encoder
if name == "visual.class_embedding":
snake_case_ :Any = name.replace("""visual.class_embedding""", """vision_model.embeddings.class_embedding""" )
if name == "visual.positional_embedding":
snake_case_ :Optional[int] = name.replace("""visual.positional_embedding""", """vision_model.embeddings.position_embedding.weight""" )
if name.startswith("""visual.transformer.resblocks""" ):
snake_case_ :Union[str, Any] = name.replace("""visual.transformer.resblocks""", """vision_model.encoder.layers""" )
if "visual.conv1" in name:
snake_case_ :int = name.replace("""visual.conv1""", """vision_model.embeddings.patch_embedding""" )
if "visual.ln_pre" in name:
snake_case_ :Any = name.replace("""visual.ln_pre""", """vision_model.pre_layernorm""" )
if "visual.ln_post" in name:
snake_case_ :str = name.replace("""visual.ln_post""", """vision_model.post_layernorm""" )
if "visual.proj" in name:
snake_case_ :Union[str, Any] = name.replace("""visual.proj""", """visual_projection.weight""" )
if "text_projection" in name:
snake_case_ :Dict = name.replace("""text_projection""", """text_projection.weight""" )
# things on top
if "prompts_visual_proj" in name:
snake_case_ :List[str] = name.replace("""prompts_visual_proj""", """prompts_visual_projection""" )
if "prompts_visual_ln" in name:
snake_case_ :Dict = name.replace("""prompts_visual_ln""", """prompts_visual_layernorm""" )
# mit
if name == "mit.positional_embedding":
snake_case_ :str = name.replace("""positional""", """position""" )
if name.startswith("""mit.resblocks""" ):
snake_case_ :Dict = name.replace("""mit.resblocks""", """mit.encoder.layers""" )
# prompts generator
if name.startswith("""prompts_generator.norm""" ):
snake_case_ :Union[str, Any] = name.replace("""prompts_generator.norm""", """prompts_generator.layernorm""" )
return name
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
for key in orig_state_dict.copy().keys():
snake_case_ :Dict = orig_state_dict.pop(_lowercase )
if "attn.in_proj" in key:
snake_case_ :Optional[Any] = key.split(""".""" )
if key.startswith("""visual""" ):
snake_case_ :Any = key_split[3]
snake_case_ :Optional[Any] = config.vision_config.hidden_size
if "message_attn" in key:
if "weight" in key:
snake_case_ :str = val[
:dim, :
]
snake_case_ :Optional[int] = val[
dim : dim * 2, :
]
snake_case_ :Union[str, Any] = val[
-dim:, :
]
else:
snake_case_ :Dict = val[
:dim
]
snake_case_ :Optional[int] = val[
dim : dim * 2
]
snake_case_ :Optional[int] = val[
-dim:
]
else:
if "weight" in key:
snake_case_ :Optional[Any] = val[
:dim, :
]
snake_case_ :List[str] = val[
dim : dim * 2, :
]
snake_case_ :Dict = val[
-dim:, :
]
else:
snake_case_ :Union[str, Any] = val[:dim]
snake_case_ :Union[str, Any] = val[
dim : dim * 2
]
snake_case_ :Union[str, Any] = val[-dim:]
elif key.startswith("""mit""" ):
snake_case_ :Tuple = key_split[2]
snake_case_ :Union[str, Any] = config.vision_config.mit_hidden_size
if "weight" in key:
snake_case_ :Optional[int] = val[:dim, :]
snake_case_ :Optional[int] = val[dim : dim * 2, :]
snake_case_ :str = val[-dim:, :]
else:
snake_case_ :str = val[:dim]
snake_case_ :Any = val[dim : dim * 2]
snake_case_ :int = val[-dim:]
else:
snake_case_ :Tuple = key_split[2]
snake_case_ :Any = config.text_config.hidden_size
if "weight" in key:
snake_case_ :Dict = val[:dim, :]
snake_case_ :Dict = val[
dim : dim * 2, :
]
snake_case_ :List[str] = val[-dim:, :]
else:
snake_case_ :Any = val[:dim]
snake_case_ :Tuple = val[
dim : dim * 2
]
snake_case_ :List[str] = val[-dim:]
else:
snake_case_ :Optional[int] = rename_key(_lowercase )
if new_key_name in ["visual_projection.weight", "text_projection.weight"]:
snake_case_ :Optional[Any] = val.T
snake_case_ :Tuple = val
return orig_state_dict
def A_ ( _lowercase ):
'''simple docstring'''
if num_frames == 8:
snake_case_ :str = """eating_spaghetti_8_frames.npy"""
elif num_frames == 16:
snake_case_ :int = """eating_spaghetti.npy"""
elif num_frames == 32:
snake_case_ :List[str] = """eating_spaghetti_32_frames.npy"""
snake_case_ :int = hf_hub_download(
repo_id="""hf-internal-testing/spaghetti-video""", filename=_lowercase, repo_type="""dataset""", )
snake_case_ :Union[str, Any] = np.load(_lowercase )
return list(_lowercase )
def A_ ( _lowercase, _lowercase=None, _lowercase=False ):
'''simple docstring'''
snake_case_ :List[Any] = {
# fully supervised kinetics-400 checkpoints
"""xclip-base-patch32""": """https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_32_8.pth""",
"""xclip-base-patch32-16-frames""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_32_16.pth"""
),
"""xclip-base-patch16""": """https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_16_8.pth""",
"""xclip-base-patch16-16-frames""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_16_16.pth"""
),
"""xclip-large-patch14""": """https://drive.google.com/u/0/uc?id=1NUOImq0o5DlQTST17iIP3vG7DgmHQuCx&export=download&confirm=t&uuid=b26caedc-88e2-473e-830a-9d158b653cdb""",
"""xclip-large-patch14-16-frames""": """https://drive.google.com/u/0/uc?id=1FOYgnJc097OJ4lGwtRCCydQyVPJEOH7d&export=download&confirm=t&uuid=538fa810-e671-4050-b385-9a623f89804f""",
# fully supervised kinetics-600 checkpoints
"""xclip-base-patch16-kinetics-600""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k600_16_8.pth"""
),
"""xclip-base-patch16-kinetics-600-16-frames""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k600_16_16.pth"""
),
"""xclip-large-patch14-kinetics-600""": """https://drive.google.com/u/0/uc?id=1FV8C1INuM91sLAN4ImjzePLIlpMSihwV&export=download&confirm=t&uuid=141d4977-4a65-44ae-864f-4b0c19f838be""",
# few shot
"""xclip-base-patch16-hmdb-2-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_2.pth"""
),
"""xclip-base-patch16-hmdb-4-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_4.pth"""
),
"""xclip-base-patch16-hmdb-8-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_8.pth"""
),
"""xclip-base-patch16-hmdb-16-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_16.pth"""
),
"""xclip-base-patch16-ucf-2-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_2.pth"""
),
"""xclip-base-patch16-ucf-4-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_4.pth"""
),
"""xclip-base-patch16-ucf-8-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_8.pth"""
),
"""xclip-base-patch16-ucf-16-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_16.pth"""
),
# zero shot
"""xclip-base-patch16-zero-shot""": """https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/zero.pth""",
}
snake_case_ :Optional[int] = model_to_url[model_name]
snake_case_ :int = 8
if "16-frames" in model_name:
snake_case_ :List[Any] = 16
elif "shot" in model_name:
snake_case_ :Dict = 32
snake_case_ :Optional[int] = get_xclip_config(_lowercase, _lowercase )
snake_case_ :Optional[Any] = XCLIPModel(_lowercase )
model.eval()
if "drive" in checkpoint_url:
snake_case_ :List[str] = """pytorch_model.bin"""
gdown.cached_download(_lowercase, _lowercase, quiet=_lowercase )
snake_case_ :List[Any] = torch.load(_lowercase, map_location="""cpu""" )["""model"""]
else:
snake_case_ :Tuple = torch.hub.load_state_dict_from_url(_lowercase )["""model"""]
snake_case_ :Union[str, Any] = convert_state_dict(_lowercase, _lowercase )
snake_case_ :str = XCLIPModel(_lowercase )
snake_case_, snake_case_ :Optional[int] = model.load_state_dict(_lowercase, strict=_lowercase )
assert missing_keys == ["text_model.embeddings.position_ids", "vision_model.embeddings.position_ids"]
model.eval()
snake_case_ :List[str] = 336 if model_name == """xclip-large-patch14-16-frames""" else 224
snake_case_ :List[Any] = VideoMAEImageProcessor(size=_lowercase )
snake_case_ :Any = CLIPTokenizer.from_pretrained("""openai/clip-vit-base-patch32""" )
snake_case_ :str = CLIPTokenizerFast.from_pretrained("""openai/clip-vit-base-patch32""" )
snake_case_ :Optional[Any] = XCLIPProcessor(image_processor=_lowercase, tokenizer=_lowercase )
snake_case_ :Optional[int] = prepare_video(_lowercase )
snake_case_ :Optional[Any] = processor(
text=["""playing sports""", """eating spaghetti""", """go shopping"""], videos=_lowercase, return_tensors="""pt""", padding=_lowercase )
print("""Shape of pixel values:""", inputs.pixel_values.shape )
with torch.no_grad():
snake_case_ :List[Any] = model(**_lowercase )
# Verify outputs
snake_case_ :List[Any] = outputs.logits_per_video
snake_case_ :Any = logits_per_video.softmax(dim=1 )
print("""Probs:""", _lowercase )
# kinetics-400
if model_name == "xclip-base-patch32":
snake_case_ :Union[str, Any] = torch.tensor([[0.0019, 0.9951, 0.0030]] )
elif model_name == "xclip-base-patch32-16-frames":
snake_case_ :str = torch.tensor([[7.09_99e-04, 9.98_83e-01, 4.55_80e-04]] )
elif model_name == "xclip-base-patch16":
snake_case_ :Tuple = torch.tensor([[0.0083, 0.9681, 0.0236]] )
elif model_name == "xclip-base-patch16-16-frames":
snake_case_ :Any = torch.tensor([[7.69_37e-04, 9.97_28e-01, 1.94_73e-03]] )
elif model_name == "xclip-large-patch14":
snake_case_ :str = torch.tensor([[0.0062, 0.9864, 0.0075]] )
elif model_name == "xclip-large-patch14-16-frames":
snake_case_ :Tuple = torch.tensor([[3.38_77e-04, 9.99_37e-01, 2.88_88e-04]] )
# kinetics-600
elif model_name == "xclip-base-patch16-kinetics-600":
snake_case_ :List[Any] = torch.tensor([[0.0555, 0.8914, 0.0531]] )
elif model_name == "xclip-base-patch16-kinetics-600-16-frames":
snake_case_ :Union[str, Any] = torch.tensor([[3.85_54e-04, 9.99_29e-01, 3.27_54e-04]] )
elif model_name == "xclip-large-patch14-kinetics-600":
snake_case_ :List[Any] = torch.tensor([[0.0036, 0.9920, 0.0045]] )
# few shot
elif model_name == "xclip-base-patch16-hmdb-2-shot":
snake_case_ :Dict = torch.tensor([[7.18_90e-06, 9.99_94e-01, 5.65_59e-05]] )
elif model_name == "xclip-base-patch16-hmdb-4-shot":
snake_case_ :Union[str, Any] = torch.tensor([[1.03_20e-05, 9.99_93e-01, 6.24_35e-05]] )
elif model_name == "xclip-base-patch16-hmdb-8-shot":
snake_case_ :str = torch.tensor([[4.13_77e-06, 9.99_90e-01, 9.83_86e-05]] )
elif model_name == "xclip-base-patch16-hmdb-16-shot":
snake_case_ :str = torch.tensor([[4.13_47e-05, 9.99_62e-01, 3.34_11e-04]] )
elif model_name == "xclip-base-patch16-ucf-2-shot":
snake_case_ :int = torch.tensor([[8.58_57e-05, 9.99_28e-01, 6.32_91e-04]] )
elif model_name == "xclip-base-patch16-ucf-4-shot":
snake_case_ :Optional[int] = torch.tensor([[8.58_57e-05, 9.99_28e-01, 6.32_91e-04]] )
elif model_name == "xclip-base-patch16-ucf-8-shot":
snake_case_ :Any = torch.tensor([[0.0027, 0.9904, 0.0070]] )
elif model_name == "xclip-base-patch16-ucf-16-shot":
snake_case_ :Tuple = torch.tensor([[9.82_19e-04, 9.95_93e-01, 3.08_63e-03]] )
# zero shot
elif model_name == "xclip-base-patch16-zero-shot":
snake_case_ :Union[str, Any] = torch.tensor([[3.50_82e-04, 9.97_85e-01, 1.79_66e-03]] )
else:
raise ValueError(f"""Model name {model_name} not supported""" )
assert torch.allclose(_lowercase, _lowercase, atol=1e-3 )
print("""Looks ok!""" )
if pytorch_dump_folder_path is not None:
print(f"""Saving model {model_name} to {pytorch_dump_folder_path}""" )
model.save_pretrained(_lowercase )
if push_to_hub:
print("""Pushing model, processor and slow tokenizer files to the hub...""" )
model.push_to_hub(_lowercase, organization="""nielsr""" )
processor.push_to_hub(_lowercase, organization="""nielsr""" )
slow_tokenizer.push_to_hub(_lowercase, organization="""nielsr""" )
if __name__ == "__main__":
__a = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--model_name",
default="xclip-base-patch32",
type=str,
help="Name of the model.",
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model directory."
)
parser.add_argument(
"--push_to_hub", action="store_true", help="Whether or not to push the converted model to the 🤗 hub."
)
__a = parser.parse_args()
convert_xclip_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 66 |
"""simple docstring"""
from __future__ import annotations
from collections import Counter
from random import random
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Tuple ) -> Optional[Any]:
snake_case_ :Optional[int] = {}
def lowerCAmelCase_ ( self: Dict , snake_case: str ) -> None:
snake_case_ :str = {}
def lowerCAmelCase_ ( self: Optional[int] , snake_case: str , snake_case: str , snake_case: float ) -> None:
if nodea not in self.connections:
self.add_node(snake_case )
if nodea not in self.connections:
self.add_node(snake_case )
snake_case_ :Dict = probability
def lowerCAmelCase_ ( self: List[Any] ) -> list[str]:
return list(self.connections )
def lowerCAmelCase_ ( self: Any , snake_case: str ) -> str:
snake_case_ :Optional[Any] = 0
snake_case_ :List[str] = random()
for dest in self.connections[node]:
current_probability += self.connections[node][dest]
if current_probability > random_value:
return dest
return ""
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :List[str] = MarkovChainGraphUndirectedUnweighted()
for nodea, nodea, probability in transitions:
graph.add_transition_probability(_lowercase, _lowercase, _lowercase )
snake_case_ :int = Counter(graph.get_nodes() )
snake_case_ :Optional[Any] = start
for _ in range(_lowercase ):
snake_case_ :Tuple = graph.transition(_lowercase )
visited[node] += 1
return visited
if __name__ == "__main__":
import doctest
doctest.testmod()
| 66 | 1 |
"""simple docstring"""
import argparse
import json
import os
from collections import OrderedDict
import numpy as np
import tensorflow as tf
import torch
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Union[str, Any] = os.path.join(args.tf_model_dir, """parameters.json""" )
snake_case_ :Any = json.loads(open(_lowercase ).read() )
if not params:
raise ValueError(
f"""It seems that the json file at {parameter_file} is empty. Make sure you have a correct json file.""" )
if not args.output.endswith(""".pt""" ):
snake_case_ :Optional[int] = args.output + """.pt"""
snake_case_ :List[str] = OrderedDict()
with tf.device("""/CPU:0""" ):
snake_case_ :Dict = tf.train.load_checkpoint(args.tf_model_dir )
snake_case_ :str = reader.get_variable_to_shape_map()
for key_name in shapes.keys():
snake_case_ :List[Any] = reader.get_tensor(_lowercase ).astype(np.floataa )
if key_name.endswith("""/adam_m""" ) or key_name.endswith("""/adam_v""" ):
continue
if key_name.startswith("""pasts/""" ):
if key_name.startswith("""pasts/mlp""" ):
snake_case_ :Any = int(key_name[9] )
elif key_name.startswith("""pasts/out""" ):
snake_case_ :Optional[int] = 8
snake_case_ :List[str] = """model.sqout.%d.weight""" % (player * 2) # enter to nn.Sequencial with Tanh, so 2 at a time
snake_case_ :Optional[Any] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :List[str] = torch.tensor(_lowercase )
elif key_name.startswith("""model/moe""" ):
snake_case_ :Tuple = int(key_name[9:].split("""/""" )[0] )
if key_name.endswith("""/switch_gating/kernel""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.feed_forward.mlp.router.classifier.weight""" % player
snake_case_ :Optional[Any] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/softmlp/kernel""" ):
snake_case_ :List[Any] = """model.blocks.%d.feed_forward.soft_bypass_mlp.weight""" % player
snake_case_ :Optional[int] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/wo/kernel""" ) or key_name.endswith("""/wi/kernel""" ):
snake_case_ :Dict = key_name[-9:-7]
for i in range(16 ):
snake_case_ :str = """model.blocks.%d.feed_forward.mlp.experts.expert_%d.%s.weight""" % (player, i, nlayer)
snake_case_ :Tuple = (
vnp[i].transpose([1, 0] ).copy()
) # In Mesh-Tensorflow, it is one array, so it is divided
snake_case_ :Optional[int] = torch.tensor(_lowercase )
elif key_name.startswith("""model/mlp""" ):
snake_case_ :Optional[int] = int(key_name[9:].split("""/""" )[0] )
if key_name.endswith("""/p1/kernel""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.feed_forward.mlp.wi.weight""" % player
snake_case_ :Dict = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/p1/bias""" ):
snake_case_ :List[Any] = """model.blocks.%d.feed_forward.mlp.wi.bias""" % player
snake_case_ :str = vnp.copy() # same because it is one dimensional
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/p2/kernel""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.feed_forward.mlp.wo.weight""" % player
snake_case_ :Dict = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Tuple = torch.tensor(_lowercase )
elif key_name.endswith("""/p2/bias""" ):
snake_case_ :Dict = """model.blocks.%d.feed_forward.mlp.wo.bias""" % player
snake_case_ :Any = vnp.copy() # same because it is one dimensional
snake_case_ :Optional[int] = torch.tensor(_lowercase )
elif key_name.startswith("""model/ln""" ):
snake_case_ :Union[str, Any] = int(key_name[8:].split("""/""" )[0] )
if key_name.endswith("""/b""" ):
snake_case_ :str = """model.blocks.%d.feed_forward.norm.bias""" % player
snake_case_ :Dict = vnp.copy() # same because it is one dimensional
snake_case_ :int = torch.tensor(_lowercase )
elif key_name.endswith("""/g""" ):
snake_case_ :Dict = """model.blocks.%d.feed_forward.norm.weight""" % player
snake_case_ :Dict = vnp.copy() # same because it is one dimensional
snake_case_ :Tuple = torch.tensor(_lowercase )
elif key_name.startswith("""model/att""" ):
snake_case_ :List[str] = int(key_name[9:].split("""/""" )[0] )
if key_name.endswith("""/qkv/kernel""" ):
snake_case_ :Optional[int] = vnp.copy() # Compute same dimension as Mesh-tensorflow using einsum
snake_case_ :Dict = state[:, 0, :, :]
snake_case_ :int = state[:, 1, :, :]
snake_case_ :List[str] = state[:, 2, :, :]
snake_case_ :str = (
state_q.reshape([state_q.shape[0], state_q.shape[1] * state_q.shape[2]] )
.transpose([1, 0] )
.copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Any = (
state_k.reshape([state_k.shape[0], state_k.shape[1] * state_k.shape[2]] )
.transpose([1, 0] )
.copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[int] = (
state_v.reshape([state_v.shape[0], state_v.shape[1] * state_v.shape[2]] )
.transpose([1, 0] )
.copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :int = """model.blocks.%d.self_attn.self_attn.q_proj.weight""" % player
snake_case_ :int = torch.tensor(_lowercase )
snake_case_ :Optional[Any] = """model.blocks.%d.self_attn.self_attn.k_proj.weight""" % player
snake_case_ :Dict = torch.tensor(_lowercase )
snake_case_ :Dict = """model.blocks.%d.self_attn.self_attn.v_proj.weight""" % player
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/o/kernel""" ):
snake_case_ :str = """model.blocks.%d.self_attn.self_attn.out_proj.weight""" % player
snake_case_ :str = (
vnp.reshape([vnp.shape[0] * vnp.shape[1], vnp.shape[2]] ).transpose([1, 0] ).copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Any = torch.tensor(_lowercase )
elif key_name.startswith("""model/an""" ):
snake_case_ :Optional[int] = int(key_name[8:].split("""/""" )[0] )
if key_name.endswith("""/b""" ):
snake_case_ :Any = """model.blocks.%d.self_attn.norm.bias""" % player
snake_case_ :Optional[int] = vnp.copy() # same because it is one dimensional
snake_case_ :Tuple = torch.tensor(_lowercase )
elif key_name.endswith("""/g""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.self_attn.norm.weight""" % player
snake_case_ :Dict = vnp.copy() # same because it is one dimensional
snake_case_ :Optional[int] = torch.tensor(_lowercase )
elif (
key_name.startswith("""model/wte""" )
or key_name.startswith("""model/wpe""" )
or key_name.startswith("""model/ete""" )
):
snake_case_ :List[Any] = {"""wte""": """embed_tokens""", """wpe""": """position_embeddings""", """ete""": """extra_position_embeddings"""}[
key_name[-3:]
]
snake_case_ :Optional[Any] = """model.%s.weight""" % nlayer
snake_case_ :Any = vnp.copy() # same in embedded
snake_case_ :List[Any] = torch.tensor(_lowercase )
if key_name.startswith("""model/wte""" ):
snake_case_ :Tuple = """lm_head.weight"""
snake_case_ :List[str] = vnp.copy() # same in embedded
snake_case_ :List[Any] = torch.tensor(_lowercase )
elif key_name.startswith("""model/wob""" ):
snake_case_ :str = """final_logits_bias"""
snake_case_ :Any = vnp.copy() # same in embedded
snake_case_ :List[Any] = state.reshape((1, -1) )
snake_case_ :Union[str, Any] = torch.tensor(_lowercase )
elif key_name == "model/dense/kernel":
snake_case_ :str = """model.last_project.weight"""
snake_case_ :Dict = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :int = torch.tensor(_lowercase )
elif key_name == "model/dense_1/bias":
snake_case_ :Optional[int] = """model.last_project.bias"""
snake_case_ :Tuple = vnp.copy() # same because it is one dimensional
snake_case_ :Any = torch.tensor(_lowercase )
torch.save(_lowercase, args.output )
if __name__ == "__main__":
__a = argparse.ArgumentParser(
description="model converter.", formatter_class=argparse.ArgumentDefaultsHelpFormatter
)
parser.add_argument("--tf_model_dir", metavar="PATH", type=str, required=True, help="import model")
parser.add_argument("--output", metavar="PATH", type=str, required=True, help="output model")
__a = parser.parse_args()
convert_tf_gptsan_to_pt(args)
| 66 |
"""simple docstring"""
import argparse
import collections
import os
import re
import tempfile
import pandas as pd
from datasets import Dataset
from huggingface_hub import hf_hub_download, upload_folder
from transformers.utils import direct_transformers_import
# All paths are set with the intent you should run this script from the root of the repo with the command
# python utils/update_metadata.py
__a = "src/transformers"
# This is to make sure the transformers module imported is the one in the repo.
__a = direct_transformers_import(TRANSFORMERS_PATH)
# Regexes that match TF/Flax/PT model names.
__a = re.compile(r"TF(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
__a = re.compile(r"Flax(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
# Will match any TF or Flax model too so need to be in an else branch afterthe two previous regexes.
__a = re.compile(r"(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
# Fill this with tuples (pipeline_tag, model_mapping, auto_model)
__a = [
("pretraining", "MODEL_FOR_PRETRAINING_MAPPING_NAMES", "AutoModelForPreTraining"),
("feature-extraction", "MODEL_MAPPING_NAMES", "AutoModel"),
("audio-classification", "MODEL_FOR_AUDIO_CLASSIFICATION_MAPPING_NAMES", "AutoModelForAudioClassification"),
("text-generation", "MODEL_FOR_CAUSAL_LM_MAPPING_NAMES", "AutoModelForCausalLM"),
("automatic-speech-recognition", "MODEL_FOR_CTC_MAPPING_NAMES", "AutoModelForCTC"),
("image-classification", "MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES", "AutoModelForImageClassification"),
("image-segmentation", "MODEL_FOR_IMAGE_SEGMENTATION_MAPPING_NAMES", "AutoModelForImageSegmentation"),
("fill-mask", "MODEL_FOR_MASKED_LM_MAPPING_NAMES", "AutoModelForMaskedLM"),
("object-detection", "MODEL_FOR_OBJECT_DETECTION_MAPPING_NAMES", "AutoModelForObjectDetection"),
(
"zero-shot-object-detection",
"MODEL_FOR_ZERO_SHOT_OBJECT_DETECTION_MAPPING_NAMES",
"AutoModelForZeroShotObjectDetection",
),
("question-answering", "MODEL_FOR_QUESTION_ANSWERING_MAPPING_NAMES", "AutoModelForQuestionAnswering"),
("text2text-generation", "MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING_NAMES", "AutoModelForSeq2SeqLM"),
("text-classification", "MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES", "AutoModelForSequenceClassification"),
("automatic-speech-recognition", "MODEL_FOR_SPEECH_SEQ_2_SEQ_MAPPING_NAMES", "AutoModelForSpeechSeq2Seq"),
(
"table-question-answering",
"MODEL_FOR_TABLE_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForTableQuestionAnswering",
),
("token-classification", "MODEL_FOR_TOKEN_CLASSIFICATION_MAPPING_NAMES", "AutoModelForTokenClassification"),
("multiple-choice", "MODEL_FOR_MULTIPLE_CHOICE_MAPPING_NAMES", "AutoModelForMultipleChoice"),
(
"next-sentence-prediction",
"MODEL_FOR_NEXT_SENTENCE_PREDICTION_MAPPING_NAMES",
"AutoModelForNextSentencePrediction",
),
(
"audio-frame-classification",
"MODEL_FOR_AUDIO_FRAME_CLASSIFICATION_MAPPING_NAMES",
"AutoModelForAudioFrameClassification",
),
("audio-xvector", "MODEL_FOR_AUDIO_XVECTOR_MAPPING_NAMES", "AutoModelForAudioXVector"),
(
"document-question-answering",
"MODEL_FOR_DOCUMENT_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForDocumentQuestionAnswering",
),
(
"visual-question-answering",
"MODEL_FOR_VISUAL_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForVisualQuestionAnswering",
),
("image-to-text", "MODEL_FOR_FOR_VISION_2_SEQ_MAPPING_NAMES", "AutoModelForVision2Seq"),
(
"zero-shot-image-classification",
"MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING_NAMES",
"AutoModelForZeroShotImageClassification",
),
("depth-estimation", "MODEL_FOR_DEPTH_ESTIMATION_MAPPING_NAMES", "AutoModelForDepthEstimation"),
("video-classification", "MODEL_FOR_VIDEO_CLASSIFICATION_MAPPING_NAMES", "AutoModelForVideoClassification"),
("mask-generation", "MODEL_FOR_MASK_GENERATION_MAPPING_NAMES", "AutoModelForMaskGeneration"),
]
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Any = re.finditer(""".+?(?:(?<=[a-z])(?=[A-Z])|(?<=[A-Z])(?=[A-Z][a-z])|$)""", _lowercase )
return [m.group(0 ) for m in matches]
def A_ ( ):
'''simple docstring'''
snake_case_ :int = transformers_module.models.auto.configuration_auto.CONFIG_MAPPING_NAMES
snake_case_ :Dict = {
config.replace("""Config""", """""" ): model_type for model_type, config in config_maping_names.items()
}
# Dictionaries flagging if each model prefix has a backend in PT/TF/Flax.
snake_case_ :Optional[Any] = collections.defaultdict(_lowercase )
snake_case_ :int = collections.defaultdict(_lowercase )
snake_case_ :List[str] = collections.defaultdict(_lowercase )
# Let's lookup through all transformers object (once) and find if models are supported by a given backend.
for attr_name in dir(_lowercase ):
snake_case_ :int = None
if _re_tf_models.match(_lowercase ) is not None:
snake_case_ :int = tf_models
snake_case_ :List[str] = _re_tf_models.match(_lowercase ).groups()[0]
elif _re_flax_models.match(_lowercase ) is not None:
snake_case_ :List[Any] = flax_models
snake_case_ :Any = _re_flax_models.match(_lowercase ).groups()[0]
elif _re_pt_models.match(_lowercase ) is not None:
snake_case_ :Optional[Any] = pt_models
snake_case_ :int = _re_pt_models.match(_lowercase ).groups()[0]
if lookup_dict is not None:
while len(_lowercase ) > 0:
if attr_name in model_prefix_to_model_type:
snake_case_ :Optional[int] = True
break
# Try again after removing the last word in the name
snake_case_ :Optional[Any] = """""".join(camel_case_split(_lowercase )[:-1] )
snake_case_ :Optional[int] = set(list(pt_models.keys() ) + list(tf_models.keys() ) + list(flax_models.keys() ) )
snake_case_ :Optional[Any] = list(_lowercase )
all_models.sort()
snake_case_ :Optional[int] = {"""model_type""": all_models}
snake_case_ :Optional[int] = [pt_models[t] for t in all_models]
snake_case_ :Any = [tf_models[t] for t in all_models]
snake_case_ :Dict = [flax_models[t] for t in all_models]
# Now let's use the auto-mapping names to make sure
snake_case_ :Dict = {}
for t in all_models:
if t in transformers_module.models.auto.processing_auto.PROCESSOR_MAPPING_NAMES:
snake_case_ :Optional[Any] = """AutoProcessor"""
elif t in transformers_module.models.auto.tokenization_auto.TOKENIZER_MAPPING_NAMES:
snake_case_ :Tuple = """AutoTokenizer"""
elif t in transformers_module.models.auto.feature_extraction_auto.FEATURE_EXTRACTOR_MAPPING_NAMES:
snake_case_ :Tuple = """AutoFeatureExtractor"""
else:
# Default to AutoTokenizer if a model has nothing, for backward compatibility.
snake_case_ :str = """AutoTokenizer"""
snake_case_ :int = [processors[t] for t in all_models]
return pd.DataFrame(_lowercase )
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :List[Any] = [
transformers_module.models.auto.modeling_auto,
transformers_module.models.auto.modeling_tf_auto,
transformers_module.models.auto.modeling_flax_auto,
]
for pipeline_tag, model_mapping, auto_class in PIPELINE_TAGS_AND_AUTO_MODELS:
snake_case_ :Optional[int] = [model_mapping, f"""TF_{model_mapping}""", f"""FLAX_{model_mapping}"""]
snake_case_ :List[str] = [auto_class, f"""TF_{auto_class}""", f"""Flax_{auto_class}"""]
# Loop through all three frameworks
for module, cls, mapping in zip(_lowercase, _lowercase, _lowercase ):
# The type of pipeline may not exist in this framework
if not hasattr(_lowercase, _lowercase ):
continue
# First extract all model_names
snake_case_ :Tuple = []
for name in getattr(_lowercase, _lowercase ).values():
if isinstance(_lowercase, _lowercase ):
model_names.append(_lowercase )
else:
model_names.extend(list(_lowercase ) )
# Add pipeline tag and auto model class for those models
table.update({model_name: (pipeline_tag, cls) for model_name in model_names} )
return table
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :List[Any] = get_frameworks_table()
snake_case_ :str = Dataset.from_pandas(_lowercase )
snake_case_ :List[Any] = hf_hub_download(
"""huggingface/transformers-metadata""", """pipeline_tags.json""", repo_type="""dataset""", token=_lowercase )
snake_case_ :List[str] = Dataset.from_json(_lowercase )
snake_case_ :int = {
tags_dataset[i]["""model_class"""]: (tags_dataset[i]["""pipeline_tag"""], tags_dataset[i]["""auto_class"""])
for i in range(len(_lowercase ) )
}
snake_case_ :Optional[int] = update_pipeline_and_auto_class_table(_lowercase )
# Sort the model classes to avoid some nondeterministic updates to create false update commits.
snake_case_ :Tuple = sorted(table.keys() )
snake_case_ :Tuple = pd.DataFrame(
{
"""model_class""": model_classes,
"""pipeline_tag""": [table[m][0] for m in model_classes],
"""auto_class""": [table[m][1] for m in model_classes],
} )
snake_case_ :Union[str, Any] = Dataset.from_pandas(_lowercase )
with tempfile.TemporaryDirectory() as tmp_dir:
frameworks_dataset.to_json(os.path.join(_lowercase, """frameworks.json""" ) )
tags_dataset.to_json(os.path.join(_lowercase, """pipeline_tags.json""" ) )
if commit_sha is not None:
snake_case_ :Union[str, Any] = (
f"""Update with commit {commit_sha}\n\nSee: """
f"""https://github.com/huggingface/transformers/commit/{commit_sha}"""
)
else:
snake_case_ :List[Any] = """Update"""
upload_folder(
repo_id="""huggingface/transformers-metadata""", folder_path=_lowercase, repo_type="""dataset""", token=_lowercase, commit_message=_lowercase, )
def A_ ( ):
'''simple docstring'''
snake_case_ :List[Any] = {tag: cls for tag, _, cls in PIPELINE_TAGS_AND_AUTO_MODELS}
snake_case_ :Dict = transformers_module.pipelines.SUPPORTED_TASKS
snake_case_ :List[str] = []
for key in pipeline_tasks:
if key not in in_table:
snake_case_ :int = pipeline_tasks[key]["""pt"""]
if isinstance(_lowercase, (list, tuple) ):
snake_case_ :Any = model[0]
snake_case_ :str = model.__name__
if model not in in_table.values():
missing.append(_lowercase )
if len(_lowercase ) > 0:
snake_case_ :Optional[int] = """, """.join(_lowercase )
raise ValueError(
"""The following pipeline tags are not present in the `PIPELINE_TAGS_AND_AUTO_MODELS` constant inside """
f"""`utils/update_metadata.py`: {msg}. Please add them!""" )
if __name__ == "__main__":
__a = argparse.ArgumentParser()
parser.add_argument("--token", type=str, help="The token to use to push to the transformers-metadata dataset.")
parser.add_argument("--commit_sha", type=str, help="The sha of the commit going with this update.")
parser.add_argument("--check-only", action="store_true", help="Activate to just check all pipelines are present.")
__a = parser.parse_args()
if args.check_only:
check_pipeline_tags()
else:
update_metadata(args.token, args.commit_sha)
| 66 | 1 |
"""simple docstring"""
import unittest
import torch
from torch import nn
from diffusers.models.activations import get_activation
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: Optional[Any] ) -> Union[str, Any]:
snake_case_ :List[str] = get_activation("""swish""" )
self.assertIsInstance(snake_case , nn.SiLU )
self.assertEqual(act(torch.tensor(-100 , dtype=torch.floataa ) ).item() , 0 )
self.assertNotEqual(act(torch.tensor(-1 , dtype=torch.floataa ) ).item() , 0 )
self.assertEqual(act(torch.tensor(0 , dtype=torch.floataa ) ).item() , 0 )
self.assertEqual(act(torch.tensor(20 , dtype=torch.floataa ) ).item() , 20 )
def lowerCAmelCase_ ( self: Optional[Any] ) -> Optional[Any]:
snake_case_ :Optional[int] = get_activation("""silu""" )
self.assertIsInstance(snake_case , nn.SiLU )
self.assertEqual(act(torch.tensor(-100 , dtype=torch.floataa ) ).item() , 0 )
self.assertNotEqual(act(torch.tensor(-1 , dtype=torch.floataa ) ).item() , 0 )
self.assertEqual(act(torch.tensor(0 , dtype=torch.floataa ) ).item() , 0 )
self.assertEqual(act(torch.tensor(20 , dtype=torch.floataa ) ).item() , 20 )
def lowerCAmelCase_ ( self: Optional[Any] ) -> Any:
snake_case_ :Optional[Any] = get_activation("""mish""" )
self.assertIsInstance(snake_case , nn.Mish )
self.assertEqual(act(torch.tensor(-200 , dtype=torch.floataa ) ).item() , 0 )
self.assertNotEqual(act(torch.tensor(-1 , dtype=torch.floataa ) ).item() , 0 )
self.assertEqual(act(torch.tensor(0 , dtype=torch.floataa ) ).item() , 0 )
self.assertEqual(act(torch.tensor(20 , dtype=torch.floataa ) ).item() , 20 )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Optional[Any]:
snake_case_ :List[Any] = get_activation("""gelu""" )
self.assertIsInstance(snake_case , nn.GELU )
self.assertEqual(act(torch.tensor(-100 , dtype=torch.floataa ) ).item() , 0 )
self.assertNotEqual(act(torch.tensor(-1 , dtype=torch.floataa ) ).item() , 0 )
self.assertEqual(act(torch.tensor(0 , dtype=torch.floataa ) ).item() , 0 )
self.assertEqual(act(torch.tensor(20 , dtype=torch.floataa ) ).item() , 20 )
| 66 |
"""simple docstring"""
import argparse
import glob
import logging
import os
from argparse import Namespace
from importlib import import_module
import numpy as np
import torch
from lightning_base import BaseTransformer, add_generic_args, generic_train
from seqeval.metrics import accuracy_score, fa_score, precision_score, recall_score
from torch.nn import CrossEntropyLoss
from torch.utils.data import DataLoader, TensorDataset
from utils_ner import TokenClassificationTask
__a = logging.getLogger(__name__)
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
_A : Union[str, Any] = """token-classification"""
def __init__( self: Any , snake_case: Tuple ) -> List[Any]:
if type(snake_case ) == dict:
snake_case_ :Optional[int] = Namespace(**snake_case )
snake_case_ :Optional[int] = import_module("""tasks""" )
try:
snake_case_ :Any = getattr(snake_case , hparams.task_type )
snake_case_ :TokenClassificationTask = token_classification_task_clazz()
except AttributeError:
raise ValueError(
f"""Task {hparams.task_type} needs to be defined as a TokenClassificationTask subclass in {module}. """
f"""Available tasks classes are: {TokenClassificationTask.__subclasses__()}""" )
snake_case_ :Any = self.token_classification_task.get_labels(hparams.labels )
snake_case_ :str = CrossEntropyLoss().ignore_index
super().__init__(snake_case , len(self.labels ) , self.mode )
def lowerCAmelCase_ ( self: Dict , **snake_case: List[Any] ) -> Any:
return self.model(**snake_case )
def lowerCAmelCase_ ( self: str , snake_case: Tuple , snake_case: List[Any] ) -> Optional[int]:
snake_case_ :List[str] = {"""input_ids""": batch[0], """attention_mask""": batch[1], """labels""": batch[3]}
if self.config.model_type != "distilbert":
snake_case_ :List[str] = (
batch[2] if self.config.model_type in ["""bert""", """xlnet"""] else None
) # XLM and RoBERTa don"t use token_type_ids
snake_case_ :Optional[Any] = self(**snake_case )
snake_case_ :List[str] = outputs[0]
# tensorboard_logs = {"loss": loss, "rate": self.lr_scheduler.get_last_lr()[-1]}
return {"loss": loss}
def lowerCAmelCase_ ( self: int ) -> Dict:
snake_case_ :List[Any] = self.hparams
for mode in ["train", "dev", "test"]:
snake_case_ :Optional[int] = self._feature_file(snake_case )
if os.path.exists(snake_case ) and not args.overwrite_cache:
logger.info("""Loading features from cached file %s""" , snake_case )
snake_case_ :Optional[int] = torch.load(snake_case )
else:
logger.info("""Creating features from dataset file at %s""" , args.data_dir )
snake_case_ :Optional[int] = self.token_classification_task.read_examples_from_file(args.data_dir , snake_case )
snake_case_ :Any = self.token_classification_task.convert_examples_to_features(
snake_case , self.labels , args.max_seq_length , self.tokenizer , cls_token_at_end=bool(self.config.model_type in ["""xlnet"""] ) , cls_token=self.tokenizer.cls_token , cls_token_segment_id=2 if self.config.model_type in ["""xlnet"""] else 0 , sep_token=self.tokenizer.sep_token , sep_token_extra=snake_case , pad_on_left=bool(self.config.model_type in ["""xlnet"""] ) , pad_token=self.tokenizer.pad_token_id , pad_token_segment_id=self.tokenizer.pad_token_type_id , pad_token_label_id=self.pad_token_label_id , )
logger.info("""Saving features into cached file %s""" , snake_case )
torch.save(snake_case , snake_case )
def lowerCAmelCase_ ( self: Optional[int] , snake_case: int , snake_case: int , snake_case: bool = False ) -> DataLoader:
snake_case_ :int = self._feature_file(snake_case )
logger.info("""Loading features from cached file %s""" , snake_case )
snake_case_ :str = torch.load(snake_case )
snake_case_ :Dict = torch.tensor([f.input_ids for f in features] , dtype=torch.long )
snake_case_ :str = torch.tensor([f.attention_mask for f in features] , dtype=torch.long )
if features[0].token_type_ids is not None:
snake_case_ :List[Any] = torch.tensor([f.token_type_ids for f in features] , dtype=torch.long )
else:
snake_case_ :List[str] = torch.tensor([0 for f in features] , dtype=torch.long )
# HACK(we will not use this anymore soon)
snake_case_ :Any = torch.tensor([f.label_ids for f in features] , dtype=torch.long )
return DataLoader(
TensorDataset(snake_case , snake_case , snake_case , snake_case ) , batch_size=snake_case )
def lowerCAmelCase_ ( self: List[str] , snake_case: Dict , snake_case: Union[str, Any] ) -> List[str]:
"""Compute validation""" ""
snake_case_ :List[str] = {"""input_ids""": batch[0], """attention_mask""": batch[1], """labels""": batch[3]}
if self.config.model_type != "distilbert":
snake_case_ :Dict = (
batch[2] if self.config.model_type in ["""bert""", """xlnet"""] else None
) # XLM and RoBERTa don"t use token_type_ids
snake_case_ :Dict = self(**snake_case )
snake_case_, snake_case_ :Dict = outputs[:2]
snake_case_ :Union[str, Any] = logits.detach().cpu().numpy()
snake_case_ :List[Any] = inputs["""labels"""].detach().cpu().numpy()
return {"val_loss": tmp_eval_loss.detach().cpu(), "pred": preds, "target": out_label_ids}
def lowerCAmelCase_ ( self: List[Any] , snake_case: int ) -> Tuple:
snake_case_ :Union[str, Any] = torch.stack([x["""val_loss"""] for x in outputs] ).mean()
snake_case_ :Tuple = np.concatenate([x["""pred"""] for x in outputs] , axis=0 )
snake_case_ :Tuple = np.argmax(snake_case , axis=2 )
snake_case_ :List[str] = np.concatenate([x["""target"""] for x in outputs] , axis=0 )
snake_case_ :Optional[Any] = dict(enumerate(self.labels ) )
snake_case_ :Dict = [[] for _ in range(out_label_ids.shape[0] )]
snake_case_ :Dict = [[] for _ in range(out_label_ids.shape[0] )]
for i in range(out_label_ids.shape[0] ):
for j in range(out_label_ids.shape[1] ):
if out_label_ids[i, j] != self.pad_token_label_id:
out_label_list[i].append(label_map[out_label_ids[i][j]] )
preds_list[i].append(label_map[preds[i][j]] )
snake_case_ :str = {
"""val_loss""": val_loss_mean,
"""accuracy_score""": accuracy_score(snake_case , snake_case ),
"""precision""": precision_score(snake_case , snake_case ),
"""recall""": recall_score(snake_case , snake_case ),
"""f1""": fa_score(snake_case , snake_case ),
}
snake_case_ :List[Any] = dict(results.items() )
snake_case_ :Union[str, Any] = results
return ret, preds_list, out_label_list
def lowerCAmelCase_ ( self: Optional[Any] , snake_case: Dict ) -> Optional[Any]:
# when stable
snake_case_, snake_case_, snake_case_ :Tuple = self._eval_end(snake_case )
snake_case_ :str = ret["""log"""]
return {"val_loss": logs["val_loss"], "log": logs, "progress_bar": logs}
def lowerCAmelCase_ ( self: Tuple , snake_case: Optional[int] ) -> Any:
# updating to test_epoch_end instead of deprecated test_end
snake_case_, snake_case_, snake_case_ :Any = self._eval_end(snake_case )
# Converting to the dict required by pl
# https://github.com/PyTorchLightning/pytorch-lightning/blob/master/\
# pytorch_lightning/trainer/logging.py#L139
snake_case_ :Optional[int] = ret["""log"""]
# `val_loss` is the key returned by `self._eval_end()` but actually refers to `test_loss`
return {"avg_test_loss": logs["val_loss"], "log": logs, "progress_bar": logs}
@staticmethod
def lowerCAmelCase_ ( snake_case: Any , snake_case: int ) -> Dict:
# Add NER specific options
BaseTransformer.add_model_specific_args(snake_case , snake_case )
parser.add_argument(
"""--task_type""" , default="""NER""" , type=snake_case , help="""Task type to fine tune in training (e.g. NER, POS, etc)""" )
parser.add_argument(
"""--max_seq_length""" , default=128 , type=snake_case , help=(
"""The maximum total input sequence length after tokenization. Sequences longer """
"""than this will be truncated, sequences shorter will be padded."""
) , )
parser.add_argument(
"""--labels""" , default="""""" , type=snake_case , help="""Path to a file containing all labels. If not specified, CoNLL-2003 labels are used.""" , )
parser.add_argument(
"""--gpus""" , default=0 , type=snake_case , help="""The number of GPUs allocated for this, it is by default 0 meaning none""" , )
parser.add_argument(
"""--overwrite_cache""" , action="""store_true""" , help="""Overwrite the cached training and evaluation sets""" )
return parser
if __name__ == "__main__":
__a = argparse.ArgumentParser()
add_generic_args(parser, os.getcwd())
__a = NERTransformer.add_model_specific_args(parser, os.getcwd())
__a = parser.parse_args()
__a = NERTransformer(args)
__a = generic_train(model, args)
if args.do_predict:
# See https://github.com/huggingface/transformers/issues/3159
# pl use this default format to create a checkpoint:
# https://github.com/PyTorchLightning/pytorch-lightning/blob/master\
# /pytorch_lightning/callbacks/model_checkpoint.py#L322
__a = sorted(glob.glob(os.path.join(args.output_dir, "checkpoint-epoch=*.ckpt"), recursive=True))
__a = model.load_from_checkpoint(checkpoints[-1])
trainer.test(model)
| 66 | 1 |
"""simple docstring"""
from __future__ import annotations
from collections import Counter
from random import random
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Tuple ) -> Optional[Any]:
snake_case_ :Optional[int] = {}
def lowerCAmelCase_ ( self: Dict , snake_case: str ) -> None:
snake_case_ :str = {}
def lowerCAmelCase_ ( self: Optional[int] , snake_case: str , snake_case: str , snake_case: float ) -> None:
if nodea not in self.connections:
self.add_node(snake_case )
if nodea not in self.connections:
self.add_node(snake_case )
snake_case_ :Dict = probability
def lowerCAmelCase_ ( self: List[Any] ) -> list[str]:
return list(self.connections )
def lowerCAmelCase_ ( self: Any , snake_case: str ) -> str:
snake_case_ :Optional[Any] = 0
snake_case_ :List[str] = random()
for dest in self.connections[node]:
current_probability += self.connections[node][dest]
if current_probability > random_value:
return dest
return ""
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :List[str] = MarkovChainGraphUndirectedUnweighted()
for nodea, nodea, probability in transitions:
graph.add_transition_probability(_lowercase, _lowercase, _lowercase )
snake_case_ :int = Counter(graph.get_nodes() )
snake_case_ :Optional[Any] = start
for _ in range(_lowercase ):
snake_case_ :Tuple = graph.transition(_lowercase )
visited[node] += 1
return visited
if __name__ == "__main__":
import doctest
doctest.testmod()
| 66 |
"""simple docstring"""
from math import factorial
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Optional[int] , snake_case: Dict , snake_case: int ) -> Tuple:
snake_case_ :List[Any] = real
if isinstance(snake_case , snake_case ):
snake_case_ :Tuple = [1] * rank
else:
snake_case_ :Optional[Any] = rank
def __repr__( self: List[str] ) -> Tuple:
return (
f"""{self.real}+"""
f"""{'+'.join(str(snake_case )+'E'+str(n+1 )for n,dual in enumerate(self.duals ) )}"""
)
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
snake_case_ :Any = self.duals.copy()
while cur[-1] == 0:
cur.pop(-1 )
return Dual(self.real , snake_case )
def __add__( self: Optional[int] , snake_case: Dict ) -> List[str]:
if not isinstance(snake_case , snake_case ):
return Dual(self.real + other , self.duals )
snake_case_ :List[Any] = self.duals.copy()
snake_case_ :Tuple = other.duals.copy()
if len(snake_case ) > len(snake_case ):
o_dual.extend([1] * (len(snake_case ) - len(snake_case )) )
elif len(snake_case ) < len(snake_case ):
s_dual.extend([1] * (len(snake_case ) - len(snake_case )) )
snake_case_ :Dict = []
for i in range(len(snake_case ) ):
new_duals.append(s_dual[i] + o_dual[i] )
return Dual(self.real + other.real , snake_case )
_A : str = __add__
def __sub__( self: Tuple , snake_case: Union[str, Any] ) -> Tuple:
return self + other * -1
def __mul__( self: str , snake_case: Tuple ) -> Optional[Any]:
if not isinstance(snake_case , snake_case ):
snake_case_ :Dict = []
for i in self.duals:
new_duals.append(i * other )
return Dual(self.real * other , snake_case )
snake_case_ :int = [0] * (len(self.duals ) + len(other.duals ) + 1)
for i, item in enumerate(self.duals ):
for j, jtem in enumerate(other.duals ):
new_duals[i + j + 1] += item * jtem
for k in range(len(self.duals ) ):
new_duals[k] += self.duals[k] * other.real
for index in range(len(other.duals ) ):
new_duals[index] += other.duals[index] * self.real
return Dual(self.real * other.real , snake_case )
_A : int = __mul__
def __truediv__( self: List[str] , snake_case: List[str] ) -> List[str]:
if not isinstance(snake_case , snake_case ):
snake_case_ :Optional[Any] = []
for i in self.duals:
new_duals.append(i / other )
return Dual(self.real / other , snake_case )
raise ValueError
def __floordiv__( self: int , snake_case: List[Any] ) -> Any:
if not isinstance(snake_case , snake_case ):
snake_case_ :Optional[int] = []
for i in self.duals:
new_duals.append(i // other )
return Dual(self.real // other , snake_case )
raise ValueError
def __pow__( self: Optional[Any] , snake_case: Optional[int] ) -> List[Any]:
if n < 0 or isinstance(snake_case , snake_case ):
raise ValueError("""power must be a positive integer""" )
if n == 0:
return 1
if n == 1:
return self
snake_case_ :str = self
for _ in range(n - 1 ):
x *= self
return x
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
if not callable(_lowercase ):
raise ValueError("""differentiate() requires a function as input for func""" )
if not isinstance(_lowercase, (float, int) ):
raise ValueError("""differentiate() requires a float as input for position""" )
if not isinstance(_lowercase, _lowercase ):
raise ValueError("""differentiate() requires an int as input for order""" )
snake_case_ :Optional[Any] = Dual(_lowercase, 1 )
snake_case_ :List[Any] = func(_lowercase )
if order == 0:
return result.real
return result.duals[order - 1] * factorial(_lowercase )
if __name__ == "__main__":
import doctest
doctest.testmod()
def A_ ( _lowercase ):
'''simple docstring'''
return y**2 * y**4
print(differentiate(f, 9, 2))
| 66 | 1 |
"""simple docstring"""
def A_ ( _lowercase = 50 ):
'''simple docstring'''
snake_case_ :Dict = [1] * (length + 1)
for row_length in range(3, length + 1 ):
for block_length in range(3, row_length + 1 ):
for block_start in range(row_length - block_length ):
ways_number[row_length] += ways_number[
row_length - block_start - block_length - 1
]
ways_number[row_length] += 1
return ways_number[length]
if __name__ == "__main__":
print(F"""{solution() = }""")
| 66 |
"""simple docstring"""
from __future__ import annotations
__a = 10
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Union[str, Any] = 1
snake_case_ :List[str] = max(_lowercase )
while placement <= max_digit:
# declare and initialize empty buckets
snake_case_ :list[list] = [[] for _ in range(_lowercase )]
# split list_of_ints between the buckets
for i in list_of_ints:
snake_case_ :Any = int((i / placement) % RADIX )
buckets[tmp].append(_lowercase )
# put each buckets' contents into list_of_ints
snake_case_ :Optional[Any] = 0
for b in range(_lowercase ):
for i in buckets[b]:
snake_case_ :Union[str, Any] = i
a += 1
# move to next
placement *= RADIX
return list_of_ints
if __name__ == "__main__":
import doctest
doctest.testmod()
| 66 | 1 |
"""simple docstring"""
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Any = [0] * len(_lowercase )
snake_case_ :List[str] = []
snake_case_ :int = []
snake_case_ :Union[str, Any] = 0
for values in graph.values():
for i in values:
indegree[i] += 1
for i in range(len(_lowercase ) ):
if indegree[i] == 0:
queue.append(_lowercase )
while queue:
snake_case_ :int = queue.pop(0 )
cnt += 1
topo.append(_lowercase )
for x in graph[vertex]:
indegree[x] -= 1
if indegree[x] == 0:
queue.append(_lowercase )
if cnt != len(_lowercase ):
print("""Cycle exists""" )
else:
print(_lowercase )
# Adjacency List of Graph
__a = {0: [1, 2], 1: [3], 2: [3], 3: [4, 5], 4: [], 5: []}
topological_sort(graph)
| 66 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_sentencepiece_available,
is_tokenizers_available,
is_torch_available,
)
__a = {"configuration_reformer": ["REFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP", "ReformerConfig"]}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["ReformerTokenizer"]
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["ReformerTokenizerFast"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"REFORMER_PRETRAINED_MODEL_ARCHIVE_LIST",
"ReformerAttention",
"ReformerForMaskedLM",
"ReformerForQuestionAnswering",
"ReformerForSequenceClassification",
"ReformerLayer",
"ReformerModel",
"ReformerModelWithLMHead",
"ReformerPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_reformer import REFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP, ReformerConfig
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_reformer import ReformerTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_reformer_fast import ReformerTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_reformer import (
REFORMER_PRETRAINED_MODEL_ARCHIVE_LIST,
ReformerAttention,
ReformerForMaskedLM,
ReformerForQuestionAnswering,
ReformerForSequenceClassification,
ReformerLayer,
ReformerModel,
ReformerModelWithLMHead,
ReformerPreTrainedModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 | 1 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_torch_available,
)
__a = {
"configuration_falcon": ["FALCON_PRETRAINED_CONFIG_ARCHIVE_MAP", "FalconConfig"],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"FALCON_PRETRAINED_MODEL_ARCHIVE_LIST",
"FalconForCausalLM",
"FalconModel",
"FalconPreTrainedModel",
"FalconForSequenceClassification",
"FalconForTokenClassification",
"FalconForQuestionAnswering",
]
if TYPE_CHECKING:
from .configuration_falcon import FALCON_PRETRAINED_CONFIG_ARCHIVE_MAP, FalconConfig
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_falcon import (
FALCON_PRETRAINED_MODEL_ARCHIVE_LIST,
FalconForCausalLM,
FalconForQuestionAnswering,
FalconForSequenceClassification,
FalconForTokenClassification,
FalconModel,
FalconPreTrainedModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 |
"""simple docstring"""
import gc
import unittest
from diffusers import FlaxControlNetModel, FlaxStableDiffusionControlNetPipeline
from diffusers.utils import is_flax_available, load_image, slow
from diffusers.utils.testing_utils import require_flax
if is_flax_available():
import jax
import jax.numpy as jnp
from flax.jax_utils import replicate
from flax.training.common_utils import shard
@slow
@require_flax
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: List[Any] ) -> Any:
# clean up the VRAM after each test
super().tearDown()
gc.collect()
def lowerCAmelCase_ ( self: Tuple ) -> Any:
snake_case_, snake_case_ :List[str] = FlaxControlNetModel.from_pretrained(
"""lllyasviel/sd-controlnet-canny""" , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_, snake_case_ :Union[str, Any] = FlaxStableDiffusionControlNetPipeline.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" , controlnet=snake_case , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_ :Union[str, Any] = controlnet_params
snake_case_ :Union[str, Any] = """bird"""
snake_case_ :List[Any] = jax.device_count()
snake_case_ :List[Any] = pipe.prepare_text_inputs([prompts] * num_samples )
snake_case_ :List[str] = load_image(
"""https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/sd_controlnet/bird_canny.png""" )
snake_case_ :List[str] = pipe.prepare_image_inputs([canny_image] * num_samples )
snake_case_ :Any = jax.random.PRNGKey(0 )
snake_case_ :List[str] = jax.random.split(snake_case , jax.device_count() )
snake_case_ :List[Any] = replicate(snake_case )
snake_case_ :List[str] = shard(snake_case )
snake_case_ :str = shard(snake_case )
snake_case_ :Dict = pipe(
prompt_ids=snake_case , image=snake_case , params=snake_case , prng_seed=snake_case , num_inference_steps=50 , jit=snake_case , ).images
assert images.shape == (jax.device_count(), 1, 768, 512, 3)
snake_case_ :str = images.reshape((images.shape[0] * images.shape[1],) + images.shape[-3:] )
snake_case_ :Union[str, Any] = images[0, 253:256, 253:256, -1]
snake_case_ :str = jnp.asarray(jax.device_get(image_slice.flatten() ) )
snake_case_ :Dict = jnp.array(
[0.1_6_7_9_6_9, 0.1_1_6_6_9_9, 0.0_8_1_5_4_3, 0.1_5_4_2_9_7, 0.1_3_2_8_1_2, 0.1_0_8_8_8_7, 0.1_6_9_9_2_2, 0.1_6_9_9_2_2, 0.2_0_5_0_7_8] )
print(f"""output_slice: {output_slice}""" )
assert jnp.abs(output_slice - expected_slice ).max() < 1E-2
def lowerCAmelCase_ ( self: int ) -> Dict:
snake_case_, snake_case_ :List[Any] = FlaxControlNetModel.from_pretrained(
"""lllyasviel/sd-controlnet-openpose""" , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_, snake_case_ :int = FlaxStableDiffusionControlNetPipeline.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" , controlnet=snake_case , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_ :str = controlnet_params
snake_case_ :Optional[int] = """Chef in the kitchen"""
snake_case_ :Union[str, Any] = jax.device_count()
snake_case_ :Any = pipe.prepare_text_inputs([prompts] * num_samples )
snake_case_ :str = load_image(
"""https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/sd_controlnet/pose.png""" )
snake_case_ :Optional[Any] = pipe.prepare_image_inputs([pose_image] * num_samples )
snake_case_ :str = jax.random.PRNGKey(0 )
snake_case_ :str = jax.random.split(snake_case , jax.device_count() )
snake_case_ :Tuple = replicate(snake_case )
snake_case_ :str = shard(snake_case )
snake_case_ :int = shard(snake_case )
snake_case_ :List[str] = pipe(
prompt_ids=snake_case , image=snake_case , params=snake_case , prng_seed=snake_case , num_inference_steps=50 , jit=snake_case , ).images
assert images.shape == (jax.device_count(), 1, 768, 512, 3)
snake_case_ :str = images.reshape((images.shape[0] * images.shape[1],) + images.shape[-3:] )
snake_case_ :int = images[0, 253:256, 253:256, -1]
snake_case_ :Dict = jnp.asarray(jax.device_get(image_slice.flatten() ) )
snake_case_ :Optional[int] = jnp.array(
[[0.2_7_1_4_8_4, 0.2_6_1_7_1_9, 0.2_7_5_3_9_1, 0.2_7_7_3_4_4, 0.2_7_9_2_9_7, 0.2_9_1_0_1_6, 0.2_9_4_9_2_2, 0.3_0_2_7_3_4, 0.3_0_2_7_3_4]] )
print(f"""output_slice: {output_slice}""" )
assert jnp.abs(output_slice - expected_slice ).max() < 1E-2
| 66 | 1 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available
__a = {
"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:
__a = [
"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
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
__a = {
"configuration_mobilebert": [
"MOBILEBERT_PRETRAINED_CONFIG_ARCHIVE_MAP",
"MobileBertConfig",
"MobileBertOnnxConfig",
],
"tokenization_mobilebert": ["MobileBertTokenizer"],
}
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["MobileBertTokenizerFast"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST",
"MobileBertForMaskedLM",
"MobileBertForMultipleChoice",
"MobileBertForNextSentencePrediction",
"MobileBertForPreTraining",
"MobileBertForQuestionAnswering",
"MobileBertForSequenceClassification",
"MobileBertForTokenClassification",
"MobileBertLayer",
"MobileBertModel",
"MobileBertPreTrainedModel",
"load_tf_weights_in_mobilebert",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"TF_MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST",
"TFMobileBertForMaskedLM",
"TFMobileBertForMultipleChoice",
"TFMobileBertForNextSentencePrediction",
"TFMobileBertForPreTraining",
"TFMobileBertForQuestionAnswering",
"TFMobileBertForSequenceClassification",
"TFMobileBertForTokenClassification",
"TFMobileBertMainLayer",
"TFMobileBertModel",
"TFMobileBertPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_mobilebert import (
MOBILEBERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
MobileBertConfig,
MobileBertOnnxConfig,
)
from .tokenization_mobilebert import MobileBertTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_mobilebert_fast import MobileBertTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_mobilebert import (
MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST,
MobileBertForMaskedLM,
MobileBertForMultipleChoice,
MobileBertForNextSentencePrediction,
MobileBertForPreTraining,
MobileBertForQuestionAnswering,
MobileBertForSequenceClassification,
MobileBertForTokenClassification,
MobileBertLayer,
MobileBertModel,
MobileBertPreTrainedModel,
load_tf_weights_in_mobilebert,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_mobilebert import (
TF_MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST,
TFMobileBertForMaskedLM,
TFMobileBertForMultipleChoice,
TFMobileBertForNextSentencePrediction,
TFMobileBertForPreTraining,
TFMobileBertForQuestionAnswering,
TFMobileBertForSequenceClassification,
TFMobileBertForTokenClassification,
TFMobileBertMainLayer,
TFMobileBertModel,
TFMobileBertPreTrainedModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 | 1 |
"""simple docstring"""
import argparse
import glob
import logging
import os
from argparse import Namespace
from importlib import import_module
import numpy as np
import torch
from lightning_base import BaseTransformer, add_generic_args, generic_train
from seqeval.metrics import accuracy_score, fa_score, precision_score, recall_score
from torch.nn import CrossEntropyLoss
from torch.utils.data import DataLoader, TensorDataset
from utils_ner import TokenClassificationTask
__a = logging.getLogger(__name__)
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
_A : Union[str, Any] = """token-classification"""
def __init__( self: Any , snake_case: Tuple ) -> List[Any]:
if type(snake_case ) == dict:
snake_case_ :Optional[int] = Namespace(**snake_case )
snake_case_ :Optional[int] = import_module("""tasks""" )
try:
snake_case_ :Any = getattr(snake_case , hparams.task_type )
snake_case_ :TokenClassificationTask = token_classification_task_clazz()
except AttributeError:
raise ValueError(
f"""Task {hparams.task_type} needs to be defined as a TokenClassificationTask subclass in {module}. """
f"""Available tasks classes are: {TokenClassificationTask.__subclasses__()}""" )
snake_case_ :Any = self.token_classification_task.get_labels(hparams.labels )
snake_case_ :str = CrossEntropyLoss().ignore_index
super().__init__(snake_case , len(self.labels ) , self.mode )
def lowerCAmelCase_ ( self: Dict , **snake_case: List[Any] ) -> Any:
return self.model(**snake_case )
def lowerCAmelCase_ ( self: str , snake_case: Tuple , snake_case: List[Any] ) -> Optional[int]:
snake_case_ :List[str] = {"""input_ids""": batch[0], """attention_mask""": batch[1], """labels""": batch[3]}
if self.config.model_type != "distilbert":
snake_case_ :List[str] = (
batch[2] if self.config.model_type in ["""bert""", """xlnet"""] else None
) # XLM and RoBERTa don"t use token_type_ids
snake_case_ :Optional[Any] = self(**snake_case )
snake_case_ :List[str] = outputs[0]
# tensorboard_logs = {"loss": loss, "rate": self.lr_scheduler.get_last_lr()[-1]}
return {"loss": loss}
def lowerCAmelCase_ ( self: int ) -> Dict:
snake_case_ :List[Any] = self.hparams
for mode in ["train", "dev", "test"]:
snake_case_ :Optional[int] = self._feature_file(snake_case )
if os.path.exists(snake_case ) and not args.overwrite_cache:
logger.info("""Loading features from cached file %s""" , snake_case )
snake_case_ :Optional[int] = torch.load(snake_case )
else:
logger.info("""Creating features from dataset file at %s""" , args.data_dir )
snake_case_ :Optional[int] = self.token_classification_task.read_examples_from_file(args.data_dir , snake_case )
snake_case_ :Any = self.token_classification_task.convert_examples_to_features(
snake_case , self.labels , args.max_seq_length , self.tokenizer , cls_token_at_end=bool(self.config.model_type in ["""xlnet"""] ) , cls_token=self.tokenizer.cls_token , cls_token_segment_id=2 if self.config.model_type in ["""xlnet"""] else 0 , sep_token=self.tokenizer.sep_token , sep_token_extra=snake_case , pad_on_left=bool(self.config.model_type in ["""xlnet"""] ) , pad_token=self.tokenizer.pad_token_id , pad_token_segment_id=self.tokenizer.pad_token_type_id , pad_token_label_id=self.pad_token_label_id , )
logger.info("""Saving features into cached file %s""" , snake_case )
torch.save(snake_case , snake_case )
def lowerCAmelCase_ ( self: Optional[int] , snake_case: int , snake_case: int , snake_case: bool = False ) -> DataLoader:
snake_case_ :int = self._feature_file(snake_case )
logger.info("""Loading features from cached file %s""" , snake_case )
snake_case_ :str = torch.load(snake_case )
snake_case_ :Dict = torch.tensor([f.input_ids for f in features] , dtype=torch.long )
snake_case_ :str = torch.tensor([f.attention_mask for f in features] , dtype=torch.long )
if features[0].token_type_ids is not None:
snake_case_ :List[Any] = torch.tensor([f.token_type_ids for f in features] , dtype=torch.long )
else:
snake_case_ :List[str] = torch.tensor([0 for f in features] , dtype=torch.long )
# HACK(we will not use this anymore soon)
snake_case_ :Any = torch.tensor([f.label_ids for f in features] , dtype=torch.long )
return DataLoader(
TensorDataset(snake_case , snake_case , snake_case , snake_case ) , batch_size=snake_case )
def lowerCAmelCase_ ( self: List[str] , snake_case: Dict , snake_case: Union[str, Any] ) -> List[str]:
"""Compute validation""" ""
snake_case_ :List[str] = {"""input_ids""": batch[0], """attention_mask""": batch[1], """labels""": batch[3]}
if self.config.model_type != "distilbert":
snake_case_ :Dict = (
batch[2] if self.config.model_type in ["""bert""", """xlnet"""] else None
) # XLM and RoBERTa don"t use token_type_ids
snake_case_ :Dict = self(**snake_case )
snake_case_, snake_case_ :Dict = outputs[:2]
snake_case_ :Union[str, Any] = logits.detach().cpu().numpy()
snake_case_ :List[Any] = inputs["""labels"""].detach().cpu().numpy()
return {"val_loss": tmp_eval_loss.detach().cpu(), "pred": preds, "target": out_label_ids}
def lowerCAmelCase_ ( self: List[Any] , snake_case: int ) -> Tuple:
snake_case_ :Union[str, Any] = torch.stack([x["""val_loss"""] for x in outputs] ).mean()
snake_case_ :Tuple = np.concatenate([x["""pred"""] for x in outputs] , axis=0 )
snake_case_ :Tuple = np.argmax(snake_case , axis=2 )
snake_case_ :List[str] = np.concatenate([x["""target"""] for x in outputs] , axis=0 )
snake_case_ :Optional[Any] = dict(enumerate(self.labels ) )
snake_case_ :Dict = [[] for _ in range(out_label_ids.shape[0] )]
snake_case_ :Dict = [[] for _ in range(out_label_ids.shape[0] )]
for i in range(out_label_ids.shape[0] ):
for j in range(out_label_ids.shape[1] ):
if out_label_ids[i, j] != self.pad_token_label_id:
out_label_list[i].append(label_map[out_label_ids[i][j]] )
preds_list[i].append(label_map[preds[i][j]] )
snake_case_ :str = {
"""val_loss""": val_loss_mean,
"""accuracy_score""": accuracy_score(snake_case , snake_case ),
"""precision""": precision_score(snake_case , snake_case ),
"""recall""": recall_score(snake_case , snake_case ),
"""f1""": fa_score(snake_case , snake_case ),
}
snake_case_ :List[Any] = dict(results.items() )
snake_case_ :Union[str, Any] = results
return ret, preds_list, out_label_list
def lowerCAmelCase_ ( self: Optional[Any] , snake_case: Dict ) -> Optional[Any]:
# when stable
snake_case_, snake_case_, snake_case_ :Tuple = self._eval_end(snake_case )
snake_case_ :str = ret["""log"""]
return {"val_loss": logs["val_loss"], "log": logs, "progress_bar": logs}
def lowerCAmelCase_ ( self: Tuple , snake_case: Optional[int] ) -> Any:
# updating to test_epoch_end instead of deprecated test_end
snake_case_, snake_case_, snake_case_ :Any = self._eval_end(snake_case )
# Converting to the dict required by pl
# https://github.com/PyTorchLightning/pytorch-lightning/blob/master/\
# pytorch_lightning/trainer/logging.py#L139
snake_case_ :Optional[int] = ret["""log"""]
# `val_loss` is the key returned by `self._eval_end()` but actually refers to `test_loss`
return {"avg_test_loss": logs["val_loss"], "log": logs, "progress_bar": logs}
@staticmethod
def lowerCAmelCase_ ( snake_case: Any , snake_case: int ) -> Dict:
# Add NER specific options
BaseTransformer.add_model_specific_args(snake_case , snake_case )
parser.add_argument(
"""--task_type""" , default="""NER""" , type=snake_case , help="""Task type to fine tune in training (e.g. NER, POS, etc)""" )
parser.add_argument(
"""--max_seq_length""" , default=128 , type=snake_case , help=(
"""The maximum total input sequence length after tokenization. Sequences longer """
"""than this will be truncated, sequences shorter will be padded."""
) , )
parser.add_argument(
"""--labels""" , default="""""" , type=snake_case , help="""Path to a file containing all labels. If not specified, CoNLL-2003 labels are used.""" , )
parser.add_argument(
"""--gpus""" , default=0 , type=snake_case , help="""The number of GPUs allocated for this, it is by default 0 meaning none""" , )
parser.add_argument(
"""--overwrite_cache""" , action="""store_true""" , help="""Overwrite the cached training and evaluation sets""" )
return parser
if __name__ == "__main__":
__a = argparse.ArgumentParser()
add_generic_args(parser, os.getcwd())
__a = NERTransformer.add_model_specific_args(parser, os.getcwd())
__a = parser.parse_args()
__a = NERTransformer(args)
__a = generic_train(model, args)
if args.do_predict:
# See https://github.com/huggingface/transformers/issues/3159
# pl use this default format to create a checkpoint:
# https://github.com/PyTorchLightning/pytorch-lightning/blob/master\
# /pytorch_lightning/callbacks/model_checkpoint.py#L322
__a = sorted(glob.glob(os.path.join(args.output_dir, "checkpoint-epoch=*.ckpt"), recursive=True))
__a = model.load_from_checkpoint(checkpoints[-1])
trainer.test(model)
| 66 |
"""simple docstring"""
import argparse
import json
import os
from collections import OrderedDict
import numpy as np
import tensorflow as tf
import torch
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Union[str, Any] = os.path.join(args.tf_model_dir, """parameters.json""" )
snake_case_ :Any = json.loads(open(_lowercase ).read() )
if not params:
raise ValueError(
f"""It seems that the json file at {parameter_file} is empty. Make sure you have a correct json file.""" )
if not args.output.endswith(""".pt""" ):
snake_case_ :Optional[int] = args.output + """.pt"""
snake_case_ :List[str] = OrderedDict()
with tf.device("""/CPU:0""" ):
snake_case_ :Dict = tf.train.load_checkpoint(args.tf_model_dir )
snake_case_ :str = reader.get_variable_to_shape_map()
for key_name in shapes.keys():
snake_case_ :List[Any] = reader.get_tensor(_lowercase ).astype(np.floataa )
if key_name.endswith("""/adam_m""" ) or key_name.endswith("""/adam_v""" ):
continue
if key_name.startswith("""pasts/""" ):
if key_name.startswith("""pasts/mlp""" ):
snake_case_ :Any = int(key_name[9] )
elif key_name.startswith("""pasts/out""" ):
snake_case_ :Optional[int] = 8
snake_case_ :List[str] = """model.sqout.%d.weight""" % (player * 2) # enter to nn.Sequencial with Tanh, so 2 at a time
snake_case_ :Optional[Any] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :List[str] = torch.tensor(_lowercase )
elif key_name.startswith("""model/moe""" ):
snake_case_ :Tuple = int(key_name[9:].split("""/""" )[0] )
if key_name.endswith("""/switch_gating/kernel""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.feed_forward.mlp.router.classifier.weight""" % player
snake_case_ :Optional[Any] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/softmlp/kernel""" ):
snake_case_ :List[Any] = """model.blocks.%d.feed_forward.soft_bypass_mlp.weight""" % player
snake_case_ :Optional[int] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/wo/kernel""" ) or key_name.endswith("""/wi/kernel""" ):
snake_case_ :Dict = key_name[-9:-7]
for i in range(16 ):
snake_case_ :str = """model.blocks.%d.feed_forward.mlp.experts.expert_%d.%s.weight""" % (player, i, nlayer)
snake_case_ :Tuple = (
vnp[i].transpose([1, 0] ).copy()
) # In Mesh-Tensorflow, it is one array, so it is divided
snake_case_ :Optional[int] = torch.tensor(_lowercase )
elif key_name.startswith("""model/mlp""" ):
snake_case_ :Optional[int] = int(key_name[9:].split("""/""" )[0] )
if key_name.endswith("""/p1/kernel""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.feed_forward.mlp.wi.weight""" % player
snake_case_ :Dict = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/p1/bias""" ):
snake_case_ :List[Any] = """model.blocks.%d.feed_forward.mlp.wi.bias""" % player
snake_case_ :str = vnp.copy() # same because it is one dimensional
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/p2/kernel""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.feed_forward.mlp.wo.weight""" % player
snake_case_ :Dict = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Tuple = torch.tensor(_lowercase )
elif key_name.endswith("""/p2/bias""" ):
snake_case_ :Dict = """model.blocks.%d.feed_forward.mlp.wo.bias""" % player
snake_case_ :Any = vnp.copy() # same because it is one dimensional
snake_case_ :Optional[int] = torch.tensor(_lowercase )
elif key_name.startswith("""model/ln""" ):
snake_case_ :Union[str, Any] = int(key_name[8:].split("""/""" )[0] )
if key_name.endswith("""/b""" ):
snake_case_ :str = """model.blocks.%d.feed_forward.norm.bias""" % player
snake_case_ :Dict = vnp.copy() # same because it is one dimensional
snake_case_ :int = torch.tensor(_lowercase )
elif key_name.endswith("""/g""" ):
snake_case_ :Dict = """model.blocks.%d.feed_forward.norm.weight""" % player
snake_case_ :Dict = vnp.copy() # same because it is one dimensional
snake_case_ :Tuple = torch.tensor(_lowercase )
elif key_name.startswith("""model/att""" ):
snake_case_ :List[str] = int(key_name[9:].split("""/""" )[0] )
if key_name.endswith("""/qkv/kernel""" ):
snake_case_ :Optional[int] = vnp.copy() # Compute same dimension as Mesh-tensorflow using einsum
snake_case_ :Dict = state[:, 0, :, :]
snake_case_ :int = state[:, 1, :, :]
snake_case_ :List[str] = state[:, 2, :, :]
snake_case_ :str = (
state_q.reshape([state_q.shape[0], state_q.shape[1] * state_q.shape[2]] )
.transpose([1, 0] )
.copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Any = (
state_k.reshape([state_k.shape[0], state_k.shape[1] * state_k.shape[2]] )
.transpose([1, 0] )
.copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[int] = (
state_v.reshape([state_v.shape[0], state_v.shape[1] * state_v.shape[2]] )
.transpose([1, 0] )
.copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :int = """model.blocks.%d.self_attn.self_attn.q_proj.weight""" % player
snake_case_ :int = torch.tensor(_lowercase )
snake_case_ :Optional[Any] = """model.blocks.%d.self_attn.self_attn.k_proj.weight""" % player
snake_case_ :Dict = torch.tensor(_lowercase )
snake_case_ :Dict = """model.blocks.%d.self_attn.self_attn.v_proj.weight""" % player
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/o/kernel""" ):
snake_case_ :str = """model.blocks.%d.self_attn.self_attn.out_proj.weight""" % player
snake_case_ :str = (
vnp.reshape([vnp.shape[0] * vnp.shape[1], vnp.shape[2]] ).transpose([1, 0] ).copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Any = torch.tensor(_lowercase )
elif key_name.startswith("""model/an""" ):
snake_case_ :Optional[int] = int(key_name[8:].split("""/""" )[0] )
if key_name.endswith("""/b""" ):
snake_case_ :Any = """model.blocks.%d.self_attn.norm.bias""" % player
snake_case_ :Optional[int] = vnp.copy() # same because it is one dimensional
snake_case_ :Tuple = torch.tensor(_lowercase )
elif key_name.endswith("""/g""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.self_attn.norm.weight""" % player
snake_case_ :Dict = vnp.copy() # same because it is one dimensional
snake_case_ :Optional[int] = torch.tensor(_lowercase )
elif (
key_name.startswith("""model/wte""" )
or key_name.startswith("""model/wpe""" )
or key_name.startswith("""model/ete""" )
):
snake_case_ :List[Any] = {"""wte""": """embed_tokens""", """wpe""": """position_embeddings""", """ete""": """extra_position_embeddings"""}[
key_name[-3:]
]
snake_case_ :Optional[Any] = """model.%s.weight""" % nlayer
snake_case_ :Any = vnp.copy() # same in embedded
snake_case_ :List[Any] = torch.tensor(_lowercase )
if key_name.startswith("""model/wte""" ):
snake_case_ :Tuple = """lm_head.weight"""
snake_case_ :List[str] = vnp.copy() # same in embedded
snake_case_ :List[Any] = torch.tensor(_lowercase )
elif key_name.startswith("""model/wob""" ):
snake_case_ :str = """final_logits_bias"""
snake_case_ :Any = vnp.copy() # same in embedded
snake_case_ :List[Any] = state.reshape((1, -1) )
snake_case_ :Union[str, Any] = torch.tensor(_lowercase )
elif key_name == "model/dense/kernel":
snake_case_ :str = """model.last_project.weight"""
snake_case_ :Dict = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :int = torch.tensor(_lowercase )
elif key_name == "model/dense_1/bias":
snake_case_ :Optional[int] = """model.last_project.bias"""
snake_case_ :Tuple = vnp.copy() # same because it is one dimensional
snake_case_ :Any = torch.tensor(_lowercase )
torch.save(_lowercase, args.output )
if __name__ == "__main__":
__a = argparse.ArgumentParser(
description="model converter.", formatter_class=argparse.ArgumentDefaultsHelpFormatter
)
parser.add_argument("--tf_model_dir", metavar="PATH", type=str, required=True, help="import model")
parser.add_argument("--output", metavar="PATH", type=str, required=True, help="output model")
__a = parser.parse_args()
convert_tf_gptsan_to_pt(args)
| 66 | 1 |
"""simple docstring"""
import pprint
import requests
__a = "https://zenquotes.io/api"
def A_ ( ):
'''simple docstring'''
return requests.get(API_ENDPOINT_URL + """/today""" ).json()
def A_ ( ):
'''simple docstring'''
return requests.get(API_ENDPOINT_URL + """/random""" ).json()
if __name__ == "__main__":
__a = random_quotes()
pprint.pprint(response)
| 66 |
"""simple docstring"""
import numpy as np
import pandas as pd
from sklearn.preprocessing import MinMaxScaler
from tensorflow.keras.layers import LSTM, Dense
from tensorflow.keras.models import Sequential
if __name__ == "__main__":
__a = pd.read_csv("sample_data.csv", header=None)
__a = df.shape[:1][0]
# If you're using some other dataset input the target column
__a = df.iloc[:, 1:2]
__a = actual_data.values.reshape(len_data, 1)
__a = MinMaxScaler().fit_transform(actual_data)
__a = 10
__a = 5
__a = 20
__a = len_data - periods * look_back
__a = actual_data[:division]
__a = actual_data[division - look_back :]
__a , __a = [], []
__a , __a = [], []
for i in range(0, len(train_data) - forward_days - look_back + 1):
train_x.append(train_data[i : i + look_back])
train_y.append(train_data[i + look_back : i + look_back + forward_days])
for i in range(0, len(test_data) - forward_days - look_back + 1):
test_x.append(test_data[i : i + look_back])
test_y.append(test_data[i + look_back : i + look_back + forward_days])
__a = np.array(train_x)
__a = np.array(test_x)
__a = np.array([list(i.ravel()) for i in train_y])
__a = np.array([list(i.ravel()) for i in test_y])
__a = Sequential()
model.add(LSTM(1_28, input_shape=(look_back, 1), return_sequences=True))
model.add(LSTM(64, input_shape=(1_28, 1)))
model.add(Dense(forward_days))
model.compile(loss="mean_squared_error", optimizer="adam")
__a = model.fit(
x_train, y_train, epochs=1_50, verbose=1, shuffle=True, batch_size=4
)
__a = model.predict(x_test)
| 66 | 1 |
"""simple docstring"""
__a = [
[0, 16, 13, 0, 0, 0],
[0, 0, 10, 12, 0, 0],
[0, 4, 0, 0, 14, 0],
[0, 0, 9, 0, 0, 20],
[0, 0, 0, 7, 0, 4],
[0, 0, 0, 0, 0, 0],
]
def A_ ( _lowercase, _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :int = [False] * len(_lowercase )
snake_case_ :int = [s]
snake_case_ :Tuple = True
while queue:
snake_case_ :int = queue.pop(0 )
for ind in range(len(graph[u] ) ):
if visited[ind] is False and graph[u][ind] > 0:
queue.append(_lowercase )
snake_case_ :List[Any] = True
snake_case_ :Optional[Any] = u
return visited[t]
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :Union[str, Any] = [-1] * (len(_lowercase ))
snake_case_ :Optional[int] = 0
snake_case_ :Any = []
snake_case_ :List[str] = [i[:] for i in graph] # Record original cut, copy.
while bfs(_lowercase, _lowercase, _lowercase, _lowercase ):
snake_case_ :Optional[Any] = float("""Inf""" )
snake_case_ :Tuple = sink
while s != source:
# Find the minimum value in select path
snake_case_ :Optional[int] = min(_lowercase, graph[parent[s]][s] )
snake_case_ :Dict = parent[s]
max_flow += path_flow
snake_case_ :List[str] = sink
while v != source:
snake_case_ :Optional[int] = parent[v]
graph[u][v] -= path_flow
graph[v][u] += path_flow
snake_case_ :str = parent[v]
for i in range(len(_lowercase ) ):
for j in range(len(graph[0] ) ):
if graph[i][j] == 0 and temp[i][j] > 0:
res.append((i, j) )
return res
if __name__ == "__main__":
print(mincut(test_graph, source=0, sink=5))
| 66 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tokenizers_available, is_torch_available
__a = {
"configuration_altclip": [
"ALTCLIP_PRETRAINED_CONFIG_ARCHIVE_MAP",
"AltCLIPConfig",
"AltCLIPTextConfig",
"AltCLIPVisionConfig",
],
"processing_altclip": ["AltCLIPProcessor"],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"ALTCLIP_PRETRAINED_MODEL_ARCHIVE_LIST",
"AltCLIPPreTrainedModel",
"AltCLIPModel",
"AltCLIPTextModel",
"AltCLIPVisionModel",
]
if TYPE_CHECKING:
from .configuration_altclip import (
ALTCLIP_PRETRAINED_CONFIG_ARCHIVE_MAP,
AltCLIPConfig,
AltCLIPTextConfig,
AltCLIPVisionConfig,
)
from .processing_altclip import AltCLIPProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_altclip import (
ALTCLIP_PRETRAINED_MODEL_ARCHIVE_LIST,
AltCLIPModel,
AltCLIPPreTrainedModel,
AltCLIPTextModel,
AltCLIPVisionModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 | 1 |
"""simple docstring"""
import argparse
from transformers import TaConfig, TaForConditionalGeneration, load_tf_weights_in_ta
from transformers.utils import logging
logging.set_verbosity_info()
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :str = TaConfig.from_json_file(_lowercase )
print(f"""Building PyTorch model from configuration: {config}""" )
snake_case_ :Optional[Any] = TaForConditionalGeneration(_lowercase )
# Load weights from tf checkpoint
load_tf_weights_in_ta(_lowercase, _lowercase, _lowercase )
# Save pytorch-model
print(f"""Save PyTorch model to {pytorch_dump_path}""" )
model.save_pretrained(_lowercase )
if __name__ == "__main__":
__a = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--tf_checkpoint_path", default=None, type=str, required=True, help="Path to the TensorFlow checkpoint path."
)
parser.add_argument(
"--config_file",
default=None,
type=str,
required=True,
help=(
"The config json file corresponding to the pre-trained T5 model. \nThis specifies the model architecture."
),
)
parser.add_argument(
"--pytorch_dump_path", default=None, type=str, required=True, help="Path to the output PyTorch model."
)
__a = parser.parse_args()
convert_tf_checkpoint_to_pytorch(args.tf_checkpoint_path, args.config_file, args.pytorch_dump_path)
| 66 |
"""simple docstring"""
import argparse
import gdown
import numpy as np
import torch
from huggingface_hub import hf_hub_download
from transformers import (
CLIPTokenizer,
CLIPTokenizerFast,
VideoMAEImageProcessor,
XCLIPConfig,
XCLIPModel,
XCLIPProcessor,
XCLIPTextConfig,
XCLIPVisionConfig,
)
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :int = XCLIPTextConfig()
# derive patch size from model name
snake_case_ :Union[str, Any] = model_name.find("""patch""" )
snake_case_ :List[str] = int(model_name[start_idx + len("""patch""" ) : start_idx + len("""patch""" ) + 2] )
snake_case_ :Any = XCLIPVisionConfig(patch_size=_lowercase, num_frames=_lowercase )
if "large" in model_name:
snake_case_ :Optional[Any] = 768
snake_case_ :Union[str, Any] = 3072
snake_case_ :Any = 12
snake_case_ :Any = 1024
snake_case_ :str = 4096
snake_case_ :Union[str, Any] = 16
snake_case_ :Union[str, Any] = 24
snake_case_ :Tuple = 768
snake_case_ :Any = 3072
if model_name == "xclip-large-patch14-16-frames":
snake_case_ :Any = 336
snake_case_ :Any = XCLIPConfig.from_text_vision_configs(_lowercase, _lowercase )
if "large" in model_name:
snake_case_ :List[Any] = 768
return config
def A_ ( _lowercase ):
'''simple docstring'''
if name == "token_embedding.weight":
snake_case_ :Optional[Any] = name.replace("""token_embedding.weight""", """text_model.embeddings.token_embedding.weight""" )
if name == "positional_embedding":
snake_case_ :Tuple = name.replace("""positional_embedding""", """text_model.embeddings.position_embedding.weight""" )
if "ln_1" in name:
snake_case_ :Dict = name.replace("""ln_1""", """layer_norm1""" )
if "ln_2" in name:
snake_case_ :str = name.replace("""ln_2""", """layer_norm2""" )
if "c_fc" in name:
snake_case_ :str = name.replace("""c_fc""", """fc1""" )
if "c_proj" in name:
snake_case_ :int = name.replace("""c_proj""", """fc2""" )
if name.startswith("""transformer.resblocks""" ):
snake_case_ :Union[str, Any] = name.replace("""transformer.resblocks""", """text_model.encoder.layers""" )
if "attn.out_proj" in name and "message" not in name:
snake_case_ :Union[str, Any] = name.replace("""attn.out_proj""", """self_attn.out_proj""" )
if "ln_final" in name:
snake_case_ :Union[str, Any] = name.replace("""ln_final""", """text_model.final_layer_norm""" )
# visual encoder
if name == "visual.class_embedding":
snake_case_ :Any = name.replace("""visual.class_embedding""", """vision_model.embeddings.class_embedding""" )
if name == "visual.positional_embedding":
snake_case_ :Optional[int] = name.replace("""visual.positional_embedding""", """vision_model.embeddings.position_embedding.weight""" )
if name.startswith("""visual.transformer.resblocks""" ):
snake_case_ :Union[str, Any] = name.replace("""visual.transformer.resblocks""", """vision_model.encoder.layers""" )
if "visual.conv1" in name:
snake_case_ :int = name.replace("""visual.conv1""", """vision_model.embeddings.patch_embedding""" )
if "visual.ln_pre" in name:
snake_case_ :Any = name.replace("""visual.ln_pre""", """vision_model.pre_layernorm""" )
if "visual.ln_post" in name:
snake_case_ :str = name.replace("""visual.ln_post""", """vision_model.post_layernorm""" )
if "visual.proj" in name:
snake_case_ :Union[str, Any] = name.replace("""visual.proj""", """visual_projection.weight""" )
if "text_projection" in name:
snake_case_ :Dict = name.replace("""text_projection""", """text_projection.weight""" )
# things on top
if "prompts_visual_proj" in name:
snake_case_ :List[str] = name.replace("""prompts_visual_proj""", """prompts_visual_projection""" )
if "prompts_visual_ln" in name:
snake_case_ :Dict = name.replace("""prompts_visual_ln""", """prompts_visual_layernorm""" )
# mit
if name == "mit.positional_embedding":
snake_case_ :str = name.replace("""positional""", """position""" )
if name.startswith("""mit.resblocks""" ):
snake_case_ :Dict = name.replace("""mit.resblocks""", """mit.encoder.layers""" )
# prompts generator
if name.startswith("""prompts_generator.norm""" ):
snake_case_ :Union[str, Any] = name.replace("""prompts_generator.norm""", """prompts_generator.layernorm""" )
return name
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
for key in orig_state_dict.copy().keys():
snake_case_ :Dict = orig_state_dict.pop(_lowercase )
if "attn.in_proj" in key:
snake_case_ :Optional[Any] = key.split(""".""" )
if key.startswith("""visual""" ):
snake_case_ :Any = key_split[3]
snake_case_ :Optional[Any] = config.vision_config.hidden_size
if "message_attn" in key:
if "weight" in key:
snake_case_ :str = val[
:dim, :
]
snake_case_ :Optional[int] = val[
dim : dim * 2, :
]
snake_case_ :Union[str, Any] = val[
-dim:, :
]
else:
snake_case_ :Dict = val[
:dim
]
snake_case_ :Optional[int] = val[
dim : dim * 2
]
snake_case_ :Optional[int] = val[
-dim:
]
else:
if "weight" in key:
snake_case_ :Optional[Any] = val[
:dim, :
]
snake_case_ :List[str] = val[
dim : dim * 2, :
]
snake_case_ :Dict = val[
-dim:, :
]
else:
snake_case_ :Union[str, Any] = val[:dim]
snake_case_ :Union[str, Any] = val[
dim : dim * 2
]
snake_case_ :Union[str, Any] = val[-dim:]
elif key.startswith("""mit""" ):
snake_case_ :Tuple = key_split[2]
snake_case_ :Union[str, Any] = config.vision_config.mit_hidden_size
if "weight" in key:
snake_case_ :Optional[int] = val[:dim, :]
snake_case_ :Optional[int] = val[dim : dim * 2, :]
snake_case_ :str = val[-dim:, :]
else:
snake_case_ :str = val[:dim]
snake_case_ :Any = val[dim : dim * 2]
snake_case_ :int = val[-dim:]
else:
snake_case_ :Tuple = key_split[2]
snake_case_ :Any = config.text_config.hidden_size
if "weight" in key:
snake_case_ :Dict = val[:dim, :]
snake_case_ :Dict = val[
dim : dim * 2, :
]
snake_case_ :List[str] = val[-dim:, :]
else:
snake_case_ :Any = val[:dim]
snake_case_ :Tuple = val[
dim : dim * 2
]
snake_case_ :List[str] = val[-dim:]
else:
snake_case_ :Optional[int] = rename_key(_lowercase )
if new_key_name in ["visual_projection.weight", "text_projection.weight"]:
snake_case_ :Optional[Any] = val.T
snake_case_ :Tuple = val
return orig_state_dict
def A_ ( _lowercase ):
'''simple docstring'''
if num_frames == 8:
snake_case_ :str = """eating_spaghetti_8_frames.npy"""
elif num_frames == 16:
snake_case_ :int = """eating_spaghetti.npy"""
elif num_frames == 32:
snake_case_ :List[str] = """eating_spaghetti_32_frames.npy"""
snake_case_ :int = hf_hub_download(
repo_id="""hf-internal-testing/spaghetti-video""", filename=_lowercase, repo_type="""dataset""", )
snake_case_ :Union[str, Any] = np.load(_lowercase )
return list(_lowercase )
def A_ ( _lowercase, _lowercase=None, _lowercase=False ):
'''simple docstring'''
snake_case_ :List[Any] = {
# fully supervised kinetics-400 checkpoints
"""xclip-base-patch32""": """https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_32_8.pth""",
"""xclip-base-patch32-16-frames""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_32_16.pth"""
),
"""xclip-base-patch16""": """https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_16_8.pth""",
"""xclip-base-patch16-16-frames""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_16_16.pth"""
),
"""xclip-large-patch14""": """https://drive.google.com/u/0/uc?id=1NUOImq0o5DlQTST17iIP3vG7DgmHQuCx&export=download&confirm=t&uuid=b26caedc-88e2-473e-830a-9d158b653cdb""",
"""xclip-large-patch14-16-frames""": """https://drive.google.com/u/0/uc?id=1FOYgnJc097OJ4lGwtRCCydQyVPJEOH7d&export=download&confirm=t&uuid=538fa810-e671-4050-b385-9a623f89804f""",
# fully supervised kinetics-600 checkpoints
"""xclip-base-patch16-kinetics-600""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k600_16_8.pth"""
),
"""xclip-base-patch16-kinetics-600-16-frames""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k600_16_16.pth"""
),
"""xclip-large-patch14-kinetics-600""": """https://drive.google.com/u/0/uc?id=1FV8C1INuM91sLAN4ImjzePLIlpMSihwV&export=download&confirm=t&uuid=141d4977-4a65-44ae-864f-4b0c19f838be""",
# few shot
"""xclip-base-patch16-hmdb-2-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_2.pth"""
),
"""xclip-base-patch16-hmdb-4-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_4.pth"""
),
"""xclip-base-patch16-hmdb-8-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_8.pth"""
),
"""xclip-base-patch16-hmdb-16-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_16.pth"""
),
"""xclip-base-patch16-ucf-2-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_2.pth"""
),
"""xclip-base-patch16-ucf-4-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_4.pth"""
),
"""xclip-base-patch16-ucf-8-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_8.pth"""
),
"""xclip-base-patch16-ucf-16-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_16.pth"""
),
# zero shot
"""xclip-base-patch16-zero-shot""": """https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/zero.pth""",
}
snake_case_ :Optional[int] = model_to_url[model_name]
snake_case_ :int = 8
if "16-frames" in model_name:
snake_case_ :List[Any] = 16
elif "shot" in model_name:
snake_case_ :Dict = 32
snake_case_ :Optional[int] = get_xclip_config(_lowercase, _lowercase )
snake_case_ :Optional[Any] = XCLIPModel(_lowercase )
model.eval()
if "drive" in checkpoint_url:
snake_case_ :List[str] = """pytorch_model.bin"""
gdown.cached_download(_lowercase, _lowercase, quiet=_lowercase )
snake_case_ :List[Any] = torch.load(_lowercase, map_location="""cpu""" )["""model"""]
else:
snake_case_ :Tuple = torch.hub.load_state_dict_from_url(_lowercase )["""model"""]
snake_case_ :Union[str, Any] = convert_state_dict(_lowercase, _lowercase )
snake_case_ :str = XCLIPModel(_lowercase )
snake_case_, snake_case_ :Optional[int] = model.load_state_dict(_lowercase, strict=_lowercase )
assert missing_keys == ["text_model.embeddings.position_ids", "vision_model.embeddings.position_ids"]
model.eval()
snake_case_ :List[str] = 336 if model_name == """xclip-large-patch14-16-frames""" else 224
snake_case_ :List[Any] = VideoMAEImageProcessor(size=_lowercase )
snake_case_ :Any = CLIPTokenizer.from_pretrained("""openai/clip-vit-base-patch32""" )
snake_case_ :str = CLIPTokenizerFast.from_pretrained("""openai/clip-vit-base-patch32""" )
snake_case_ :Optional[Any] = XCLIPProcessor(image_processor=_lowercase, tokenizer=_lowercase )
snake_case_ :Optional[int] = prepare_video(_lowercase )
snake_case_ :Optional[Any] = processor(
text=["""playing sports""", """eating spaghetti""", """go shopping"""], videos=_lowercase, return_tensors="""pt""", padding=_lowercase )
print("""Shape of pixel values:""", inputs.pixel_values.shape )
with torch.no_grad():
snake_case_ :List[Any] = model(**_lowercase )
# Verify outputs
snake_case_ :List[Any] = outputs.logits_per_video
snake_case_ :Any = logits_per_video.softmax(dim=1 )
print("""Probs:""", _lowercase )
# kinetics-400
if model_name == "xclip-base-patch32":
snake_case_ :Union[str, Any] = torch.tensor([[0.0019, 0.9951, 0.0030]] )
elif model_name == "xclip-base-patch32-16-frames":
snake_case_ :str = torch.tensor([[7.09_99e-04, 9.98_83e-01, 4.55_80e-04]] )
elif model_name == "xclip-base-patch16":
snake_case_ :Tuple = torch.tensor([[0.0083, 0.9681, 0.0236]] )
elif model_name == "xclip-base-patch16-16-frames":
snake_case_ :Any = torch.tensor([[7.69_37e-04, 9.97_28e-01, 1.94_73e-03]] )
elif model_name == "xclip-large-patch14":
snake_case_ :str = torch.tensor([[0.0062, 0.9864, 0.0075]] )
elif model_name == "xclip-large-patch14-16-frames":
snake_case_ :Tuple = torch.tensor([[3.38_77e-04, 9.99_37e-01, 2.88_88e-04]] )
# kinetics-600
elif model_name == "xclip-base-patch16-kinetics-600":
snake_case_ :List[Any] = torch.tensor([[0.0555, 0.8914, 0.0531]] )
elif model_name == "xclip-base-patch16-kinetics-600-16-frames":
snake_case_ :Union[str, Any] = torch.tensor([[3.85_54e-04, 9.99_29e-01, 3.27_54e-04]] )
elif model_name == "xclip-large-patch14-kinetics-600":
snake_case_ :List[Any] = torch.tensor([[0.0036, 0.9920, 0.0045]] )
# few shot
elif model_name == "xclip-base-patch16-hmdb-2-shot":
snake_case_ :Dict = torch.tensor([[7.18_90e-06, 9.99_94e-01, 5.65_59e-05]] )
elif model_name == "xclip-base-patch16-hmdb-4-shot":
snake_case_ :Union[str, Any] = torch.tensor([[1.03_20e-05, 9.99_93e-01, 6.24_35e-05]] )
elif model_name == "xclip-base-patch16-hmdb-8-shot":
snake_case_ :str = torch.tensor([[4.13_77e-06, 9.99_90e-01, 9.83_86e-05]] )
elif model_name == "xclip-base-patch16-hmdb-16-shot":
snake_case_ :str = torch.tensor([[4.13_47e-05, 9.99_62e-01, 3.34_11e-04]] )
elif model_name == "xclip-base-patch16-ucf-2-shot":
snake_case_ :int = torch.tensor([[8.58_57e-05, 9.99_28e-01, 6.32_91e-04]] )
elif model_name == "xclip-base-patch16-ucf-4-shot":
snake_case_ :Optional[int] = torch.tensor([[8.58_57e-05, 9.99_28e-01, 6.32_91e-04]] )
elif model_name == "xclip-base-patch16-ucf-8-shot":
snake_case_ :Any = torch.tensor([[0.0027, 0.9904, 0.0070]] )
elif model_name == "xclip-base-patch16-ucf-16-shot":
snake_case_ :Tuple = torch.tensor([[9.82_19e-04, 9.95_93e-01, 3.08_63e-03]] )
# zero shot
elif model_name == "xclip-base-patch16-zero-shot":
snake_case_ :Union[str, Any] = torch.tensor([[3.50_82e-04, 9.97_85e-01, 1.79_66e-03]] )
else:
raise ValueError(f"""Model name {model_name} not supported""" )
assert torch.allclose(_lowercase, _lowercase, atol=1e-3 )
print("""Looks ok!""" )
if pytorch_dump_folder_path is not None:
print(f"""Saving model {model_name} to {pytorch_dump_folder_path}""" )
model.save_pretrained(_lowercase )
if push_to_hub:
print("""Pushing model, processor and slow tokenizer files to the hub...""" )
model.push_to_hub(_lowercase, organization="""nielsr""" )
processor.push_to_hub(_lowercase, organization="""nielsr""" )
slow_tokenizer.push_to_hub(_lowercase, organization="""nielsr""" )
if __name__ == "__main__":
__a = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--model_name",
default="xclip-base-patch32",
type=str,
help="Name of the model.",
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model directory."
)
parser.add_argument(
"--push_to_hub", action="store_true", help="Whether or not to push the converted model to the 🤗 hub."
)
__a = parser.parse_args()
convert_xclip_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 66 | 1 |
"""simple docstring"""
import argparse
import collections
import os
import re
import tempfile
import pandas as pd
from datasets import Dataset
from huggingface_hub import hf_hub_download, upload_folder
from transformers.utils import direct_transformers_import
# All paths are set with the intent you should run this script from the root of the repo with the command
# python utils/update_metadata.py
__a = "src/transformers"
# This is to make sure the transformers module imported is the one in the repo.
__a = direct_transformers_import(TRANSFORMERS_PATH)
# Regexes that match TF/Flax/PT model names.
__a = re.compile(r"TF(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
__a = re.compile(r"Flax(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
# Will match any TF or Flax model too so need to be in an else branch afterthe two previous regexes.
__a = re.compile(r"(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
# Fill this with tuples (pipeline_tag, model_mapping, auto_model)
__a = [
("pretraining", "MODEL_FOR_PRETRAINING_MAPPING_NAMES", "AutoModelForPreTraining"),
("feature-extraction", "MODEL_MAPPING_NAMES", "AutoModel"),
("audio-classification", "MODEL_FOR_AUDIO_CLASSIFICATION_MAPPING_NAMES", "AutoModelForAudioClassification"),
("text-generation", "MODEL_FOR_CAUSAL_LM_MAPPING_NAMES", "AutoModelForCausalLM"),
("automatic-speech-recognition", "MODEL_FOR_CTC_MAPPING_NAMES", "AutoModelForCTC"),
("image-classification", "MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES", "AutoModelForImageClassification"),
("image-segmentation", "MODEL_FOR_IMAGE_SEGMENTATION_MAPPING_NAMES", "AutoModelForImageSegmentation"),
("fill-mask", "MODEL_FOR_MASKED_LM_MAPPING_NAMES", "AutoModelForMaskedLM"),
("object-detection", "MODEL_FOR_OBJECT_DETECTION_MAPPING_NAMES", "AutoModelForObjectDetection"),
(
"zero-shot-object-detection",
"MODEL_FOR_ZERO_SHOT_OBJECT_DETECTION_MAPPING_NAMES",
"AutoModelForZeroShotObjectDetection",
),
("question-answering", "MODEL_FOR_QUESTION_ANSWERING_MAPPING_NAMES", "AutoModelForQuestionAnswering"),
("text2text-generation", "MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING_NAMES", "AutoModelForSeq2SeqLM"),
("text-classification", "MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES", "AutoModelForSequenceClassification"),
("automatic-speech-recognition", "MODEL_FOR_SPEECH_SEQ_2_SEQ_MAPPING_NAMES", "AutoModelForSpeechSeq2Seq"),
(
"table-question-answering",
"MODEL_FOR_TABLE_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForTableQuestionAnswering",
),
("token-classification", "MODEL_FOR_TOKEN_CLASSIFICATION_MAPPING_NAMES", "AutoModelForTokenClassification"),
("multiple-choice", "MODEL_FOR_MULTIPLE_CHOICE_MAPPING_NAMES", "AutoModelForMultipleChoice"),
(
"next-sentence-prediction",
"MODEL_FOR_NEXT_SENTENCE_PREDICTION_MAPPING_NAMES",
"AutoModelForNextSentencePrediction",
),
(
"audio-frame-classification",
"MODEL_FOR_AUDIO_FRAME_CLASSIFICATION_MAPPING_NAMES",
"AutoModelForAudioFrameClassification",
),
("audio-xvector", "MODEL_FOR_AUDIO_XVECTOR_MAPPING_NAMES", "AutoModelForAudioXVector"),
(
"document-question-answering",
"MODEL_FOR_DOCUMENT_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForDocumentQuestionAnswering",
),
(
"visual-question-answering",
"MODEL_FOR_VISUAL_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForVisualQuestionAnswering",
),
("image-to-text", "MODEL_FOR_FOR_VISION_2_SEQ_MAPPING_NAMES", "AutoModelForVision2Seq"),
(
"zero-shot-image-classification",
"MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING_NAMES",
"AutoModelForZeroShotImageClassification",
),
("depth-estimation", "MODEL_FOR_DEPTH_ESTIMATION_MAPPING_NAMES", "AutoModelForDepthEstimation"),
("video-classification", "MODEL_FOR_VIDEO_CLASSIFICATION_MAPPING_NAMES", "AutoModelForVideoClassification"),
("mask-generation", "MODEL_FOR_MASK_GENERATION_MAPPING_NAMES", "AutoModelForMaskGeneration"),
]
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Any = re.finditer(""".+?(?:(?<=[a-z])(?=[A-Z])|(?<=[A-Z])(?=[A-Z][a-z])|$)""", _lowercase )
return [m.group(0 ) for m in matches]
def A_ ( ):
'''simple docstring'''
snake_case_ :int = transformers_module.models.auto.configuration_auto.CONFIG_MAPPING_NAMES
snake_case_ :Dict = {
config.replace("""Config""", """""" ): model_type for model_type, config in config_maping_names.items()
}
# Dictionaries flagging if each model prefix has a backend in PT/TF/Flax.
snake_case_ :Optional[Any] = collections.defaultdict(_lowercase )
snake_case_ :int = collections.defaultdict(_lowercase )
snake_case_ :List[str] = collections.defaultdict(_lowercase )
# Let's lookup through all transformers object (once) and find if models are supported by a given backend.
for attr_name in dir(_lowercase ):
snake_case_ :int = None
if _re_tf_models.match(_lowercase ) is not None:
snake_case_ :int = tf_models
snake_case_ :List[str] = _re_tf_models.match(_lowercase ).groups()[0]
elif _re_flax_models.match(_lowercase ) is not None:
snake_case_ :List[Any] = flax_models
snake_case_ :Any = _re_flax_models.match(_lowercase ).groups()[0]
elif _re_pt_models.match(_lowercase ) is not None:
snake_case_ :Optional[Any] = pt_models
snake_case_ :int = _re_pt_models.match(_lowercase ).groups()[0]
if lookup_dict is not None:
while len(_lowercase ) > 0:
if attr_name in model_prefix_to_model_type:
snake_case_ :Optional[int] = True
break
# Try again after removing the last word in the name
snake_case_ :Optional[Any] = """""".join(camel_case_split(_lowercase )[:-1] )
snake_case_ :Optional[int] = set(list(pt_models.keys() ) + list(tf_models.keys() ) + list(flax_models.keys() ) )
snake_case_ :Optional[Any] = list(_lowercase )
all_models.sort()
snake_case_ :Optional[int] = {"""model_type""": all_models}
snake_case_ :Optional[int] = [pt_models[t] for t in all_models]
snake_case_ :Any = [tf_models[t] for t in all_models]
snake_case_ :Dict = [flax_models[t] for t in all_models]
# Now let's use the auto-mapping names to make sure
snake_case_ :Dict = {}
for t in all_models:
if t in transformers_module.models.auto.processing_auto.PROCESSOR_MAPPING_NAMES:
snake_case_ :Optional[Any] = """AutoProcessor"""
elif t in transformers_module.models.auto.tokenization_auto.TOKENIZER_MAPPING_NAMES:
snake_case_ :Tuple = """AutoTokenizer"""
elif t in transformers_module.models.auto.feature_extraction_auto.FEATURE_EXTRACTOR_MAPPING_NAMES:
snake_case_ :Tuple = """AutoFeatureExtractor"""
else:
# Default to AutoTokenizer if a model has nothing, for backward compatibility.
snake_case_ :str = """AutoTokenizer"""
snake_case_ :int = [processors[t] for t in all_models]
return pd.DataFrame(_lowercase )
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :List[Any] = [
transformers_module.models.auto.modeling_auto,
transformers_module.models.auto.modeling_tf_auto,
transformers_module.models.auto.modeling_flax_auto,
]
for pipeline_tag, model_mapping, auto_class in PIPELINE_TAGS_AND_AUTO_MODELS:
snake_case_ :Optional[int] = [model_mapping, f"""TF_{model_mapping}""", f"""FLAX_{model_mapping}"""]
snake_case_ :List[str] = [auto_class, f"""TF_{auto_class}""", f"""Flax_{auto_class}"""]
# Loop through all three frameworks
for module, cls, mapping in zip(_lowercase, _lowercase, _lowercase ):
# The type of pipeline may not exist in this framework
if not hasattr(_lowercase, _lowercase ):
continue
# First extract all model_names
snake_case_ :Tuple = []
for name in getattr(_lowercase, _lowercase ).values():
if isinstance(_lowercase, _lowercase ):
model_names.append(_lowercase )
else:
model_names.extend(list(_lowercase ) )
# Add pipeline tag and auto model class for those models
table.update({model_name: (pipeline_tag, cls) for model_name in model_names} )
return table
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :List[Any] = get_frameworks_table()
snake_case_ :str = Dataset.from_pandas(_lowercase )
snake_case_ :List[Any] = hf_hub_download(
"""huggingface/transformers-metadata""", """pipeline_tags.json""", repo_type="""dataset""", token=_lowercase )
snake_case_ :List[str] = Dataset.from_json(_lowercase )
snake_case_ :int = {
tags_dataset[i]["""model_class"""]: (tags_dataset[i]["""pipeline_tag"""], tags_dataset[i]["""auto_class"""])
for i in range(len(_lowercase ) )
}
snake_case_ :Optional[int] = update_pipeline_and_auto_class_table(_lowercase )
# Sort the model classes to avoid some nondeterministic updates to create false update commits.
snake_case_ :Tuple = sorted(table.keys() )
snake_case_ :Tuple = pd.DataFrame(
{
"""model_class""": model_classes,
"""pipeline_tag""": [table[m][0] for m in model_classes],
"""auto_class""": [table[m][1] for m in model_classes],
} )
snake_case_ :Union[str, Any] = Dataset.from_pandas(_lowercase )
with tempfile.TemporaryDirectory() as tmp_dir:
frameworks_dataset.to_json(os.path.join(_lowercase, """frameworks.json""" ) )
tags_dataset.to_json(os.path.join(_lowercase, """pipeline_tags.json""" ) )
if commit_sha is not None:
snake_case_ :Union[str, Any] = (
f"""Update with commit {commit_sha}\n\nSee: """
f"""https://github.com/huggingface/transformers/commit/{commit_sha}"""
)
else:
snake_case_ :List[Any] = """Update"""
upload_folder(
repo_id="""huggingface/transformers-metadata""", folder_path=_lowercase, repo_type="""dataset""", token=_lowercase, commit_message=_lowercase, )
def A_ ( ):
'''simple docstring'''
snake_case_ :List[Any] = {tag: cls for tag, _, cls in PIPELINE_TAGS_AND_AUTO_MODELS}
snake_case_ :Dict = transformers_module.pipelines.SUPPORTED_TASKS
snake_case_ :List[str] = []
for key in pipeline_tasks:
if key not in in_table:
snake_case_ :int = pipeline_tasks[key]["""pt"""]
if isinstance(_lowercase, (list, tuple) ):
snake_case_ :Any = model[0]
snake_case_ :str = model.__name__
if model not in in_table.values():
missing.append(_lowercase )
if len(_lowercase ) > 0:
snake_case_ :Optional[int] = """, """.join(_lowercase )
raise ValueError(
"""The following pipeline tags are not present in the `PIPELINE_TAGS_AND_AUTO_MODELS` constant inside """
f"""`utils/update_metadata.py`: {msg}. Please add them!""" )
if __name__ == "__main__":
__a = argparse.ArgumentParser()
parser.add_argument("--token", type=str, help="The token to use to push to the transformers-metadata dataset.")
parser.add_argument("--commit_sha", type=str, help="The sha of the commit going with this update.")
parser.add_argument("--check-only", action="store_true", help="Activate to just check all pipelines are present.")
__a = parser.parse_args()
if args.check_only:
check_pipeline_tags()
else:
update_metadata(args.token, args.commit_sha)
| 66 |
"""simple docstring"""
import unittest
import numpy as np
from transformers import BertConfig, is_flax_available
from transformers.testing_utils import require_flax, slow
from ...test_modeling_flax_common import FlaxModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask
if is_flax_available():
from transformers.models.bert.modeling_flax_bert import (
FlaxBertForMaskedLM,
FlaxBertForMultipleChoice,
FlaxBertForNextSentencePrediction,
FlaxBertForPreTraining,
FlaxBertForQuestionAnswering,
FlaxBertForSequenceClassification,
FlaxBertForTokenClassification,
FlaxBertModel,
)
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def __init__( self: List[Any] , snake_case: List[str] , snake_case: Optional[Any]=13 , snake_case: List[str]=7 , snake_case: Dict=True , snake_case: List[str]=True , snake_case: Optional[int]=True , snake_case: Any=True , snake_case: Optional[Any]=99 , snake_case: Tuple=32 , snake_case: Tuple=5 , snake_case: Dict=4 , snake_case: Optional[Any]=37 , snake_case: Union[str, Any]="gelu" , snake_case: Tuple=0.1 , snake_case: List[Any]=0.1 , snake_case: List[str]=512 , snake_case: Optional[int]=16 , snake_case: int=2 , snake_case: List[Any]=0.0_2 , snake_case: Union[str, Any]=4 , ) -> List[str]:
snake_case_ :Dict = parent
snake_case_ :Any = batch_size
snake_case_ :Any = seq_length
snake_case_ :List[str] = is_training
snake_case_ :Optional[Any] = use_attention_mask
snake_case_ :Dict = use_token_type_ids
snake_case_ :Union[str, Any] = use_labels
snake_case_ :str = vocab_size
snake_case_ :int = hidden_size
snake_case_ :List[str] = num_hidden_layers
snake_case_ :Dict = num_attention_heads
snake_case_ :Any = intermediate_size
snake_case_ :Tuple = hidden_act
snake_case_ :int = hidden_dropout_prob
snake_case_ :Optional[Any] = attention_probs_dropout_prob
snake_case_ :Any = max_position_embeddings
snake_case_ :Union[str, Any] = type_vocab_size
snake_case_ :Optional[int] = type_sequence_label_size
snake_case_ :Union[str, Any] = initializer_range
snake_case_ :Tuple = num_choices
def lowerCAmelCase_ ( self: Tuple ) -> str:
snake_case_ :Optional[Any] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case_ :Union[str, Any] = None
if self.use_attention_mask:
snake_case_ :str = random_attention_mask([self.batch_size, self.seq_length] )
snake_case_ :Any = None
if self.use_token_type_ids:
snake_case_ :List[str] = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
snake_case_ :int = BertConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , is_decoder=snake_case , initializer_range=self.initializer_range , )
return config, input_ids, token_type_ids, attention_mask
def lowerCAmelCase_ ( self: Optional[int] ) -> int:
snake_case_ :str = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_, snake_case_ :Optional[int] = config_and_inputs
snake_case_ :Union[str, Any] = {"""input_ids""": input_ids, """token_type_ids""": token_type_ids, """attention_mask""": attention_mask}
return config, inputs_dict
def lowerCAmelCase_ ( self: Optional[Any] ) -> Any:
snake_case_ :int = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_, snake_case_ :Dict = config_and_inputs
snake_case_ :Union[str, Any] = True
snake_case_ :Optional[int] = floats_tensor([self.batch_size, self.seq_length, self.hidden_size] )
snake_case_ :Tuple = ids_tensor([self.batch_size, self.seq_length] , vocab_size=2 )
return (
config,
input_ids,
attention_mask,
encoder_hidden_states,
encoder_attention_mask,
)
@require_flax
class lowerCamelCase ( _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[str] = True
_A : Dict = (
(
FlaxBertModel,
FlaxBertForPreTraining,
FlaxBertForMaskedLM,
FlaxBertForMultipleChoice,
FlaxBertForQuestionAnswering,
FlaxBertForNextSentencePrediction,
FlaxBertForSequenceClassification,
FlaxBertForTokenClassification,
FlaxBertForQuestionAnswering,
)
if is_flax_available()
else ()
)
def lowerCAmelCase_ ( self: int ) -> List[str]:
snake_case_ :Any = FlaxBertModelTester(self )
@slow
def lowerCAmelCase_ ( self: List[str] ) -> Dict:
# Only check this for base model, not necessary for all model classes.
# This will also help speed-up tests.
snake_case_ :Dict = FlaxBertModel.from_pretrained("""bert-base-cased""" )
snake_case_ :Dict = model(np.ones((1, 1) ) )
self.assertIsNotNone(snake_case )
| 66 | 1 |
"""simple docstring"""
import os
from itertools import chain
from random import randrange, shuffle
import pytest
from .sola import PokerHand
__a = (
"4S 3H 2C 7S 5H",
"9D 8H 2C 6S 7H",
"2D 6D 9D TH 7D",
"TC 8C 2S JH 6C",
"JH 8S TH AH QH",
"TS KS 5S 9S AC",
"KD 6S 9D TH AD",
"KS 8D 4D 9S 4S", # pair
"8C 4S KH JS 4D", # pair
"QH 8H KD JH 8S", # pair
"KC 4H KS 2H 8D", # pair
"KD 4S KC 3H 8S", # pair
"AH 8S AS KC JH", # pair
"3H 4C 4H 3S 2H", # 2 pairs
"5S 5D 2C KH KH", # 2 pairs
"3C KH 5D 5S KH", # 2 pairs
"AS 3C KH AD KH", # 2 pairs
"7C 7S 3S 7H 5S", # 3 of a kind
"7C 7S KH 2H 7H", # 3 of a kind
"AC KH QH AH AS", # 3 of a kind
"2H 4D 3C AS 5S", # straight (low ace)
"3C 5C 4C 2C 6H", # straight
"6S 8S 7S 5H 9H", # straight
"JS QS 9H TS KH", # straight
"QC KH TS JS AH", # straight (high ace)
"8C 9C 5C 3C TC", # flush
"3S 8S 9S 5S KS", # flush
"4C 5C 9C 8C KC", # flush
"JH 8H AH KH QH", # flush
"3D 2H 3H 2C 2D", # full house
"2H 2C 3S 3H 3D", # full house
"KH KC 3S 3H 3D", # full house
"JC 6H JS JD JH", # 4 of a kind
"JC 7H JS JD JH", # 4 of a kind
"JC KH JS JD JH", # 4 of a kind
"2S AS 4S 5S 3S", # straight flush (low ace)
"2D 6D 3D 4D 5D", # straight flush
"5C 6C 3C 7C 4C", # straight flush
"JH 9H TH KH QH", # straight flush
"JH AH TH KH QH", # royal flush (high ace straight flush)
)
__a = (
("2H 3H 4H 5H 6H", "KS AS TS QS JS", "Loss"),
("2H 3H 4H 5H 6H", "AS AD AC AH JD", "Win"),
("AS AH 2H AD AC", "JS JD JC JH 3D", "Win"),
("2S AH 2H AS AC", "JS JD JC JH AD", "Loss"),
("2S AH 2H AS AC", "2H 3H 5H 6H 7H", "Win"),
("AS 3S 4S 8S 2S", "2H 3H 5H 6H 7H", "Win"),
("2H 3H 5H 6H 7H", "2S 3H 4H 5S 6C", "Win"),
("2S 3H 4H 5S 6C", "3D 4C 5H 6H 2S", "Tie"),
("2S 3H 4H 5S 6C", "AH AC 5H 6H AS", "Win"),
("2S 2H 4H 5S 4C", "AH AC 5H 6H AS", "Loss"),
("2S 2H 4H 5S 4C", "AH AC 5H 6H 7S", "Win"),
("6S AD 7H 4S AS", "AH AC 5H 6H 7S", "Loss"),
("2S AH 4H 5S KC", "AH AC 5H 6H 7S", "Loss"),
("2S 3H 6H 7S 9C", "7H 3C TH 6H 9S", "Loss"),
("4S 5H 6H TS AC", "3S 5H 6H TS AC", "Win"),
("2S AH 4H 5S 6C", "AD 4C 5H 6H 2C", "Tie"),
("AS AH 3H AD AC", "AS AH 2H AD AC", "Win"),
("AH AC 5H 5C QS", "AH AC 5H 5C KS", "Loss"),
("AH AC 5H 5C QS", "KH KC 5H 5C QS", "Win"),
("7C 7S KH 2H 7H", "3C 3S AH 2H 3H", "Win"),
("3C 3S AH 2H 3H", "7C 7S KH 2H 7H", "Loss"),
("6H 5H 4H 3H 2H", "5H 4H 3H 2H AH", "Win"),
("5H 4H 3H 2H AH", "5H 4H 3H 2H AH", "Tie"),
("5H 4H 3H 2H AH", "6H 5H 4H 3H 2H", "Loss"),
("AH AD KS KC AC", "AH KD KH AC KC", "Win"),
("2H 4D 3C AS 5S", "2H 4D 3C 6S 5S", "Loss"),
("2H 3S 3C 3H 2S", "3S 3C 2S 2H 2D", "Win"),
("4D 6D 5D 2D JH", "3S 8S 3H TC KH", "Loss"),
("4S 6C 8S 3S 7S", "AD KS 2D 7D 7C", "Loss"),
("6S 4C 7H 8C 3H", "5H JC AH 9D 9C", "Loss"),
("9D 9H JH TC QH", "3C 2S JS 5C 7H", "Win"),
("2H TC 8S AD 9S", "4H TS 7H 2C 5C", "Win"),
("9D 3S 2C 7S 7C", "JC TD 3C TC 9H", "Loss"),
)
__a = (
("2H 3H 4H 5H 6H", True),
("AS AH 2H AD AC", False),
("2H 3H 5H 6H 7H", True),
("KS AS TS QS JS", True),
("8H 9H QS JS TH", False),
("AS 3S 4S 8S 2S", True),
)
__a = (
("2H 3H 4H 5H 6H", True),
("AS AH 2H AD AC", False),
("2H 3H 5H 6H 7H", False),
("KS AS TS QS JS", True),
("8H 9H QS JS TH", True),
)
__a = (
("2H 4D 3C AS 5S", True, [5, 4, 3, 2, 14]),
("2H 5D 3C AS 5S", False, [14, 5, 5, 3, 2]),
("JH QD KC AS TS", False, [14, 13, 12, 11, 10]),
("9D 3S 2C 7S 7C", False, [9, 7, 7, 3, 2]),
)
__a = (
("JH AH TH KH QH", 0),
("JH 9H TH KH QH", 0),
("JC KH JS JD JH", 7),
("KH KC 3S 3H 3D", 6),
("8C 9C 5C 3C TC", 0),
("JS QS 9H TS KH", 0),
("7C 7S KH 2H 7H", 3),
("3C KH 5D 5S KH", 2),
("QH 8H KD JH 8S", 1),
("2D 6D 9D TH 7D", 0),
)
__a = (
("JH AH TH KH QH", 23),
("JH 9H TH KH QH", 22),
("JC KH JS JD JH", 21),
("KH KC 3S 3H 3D", 20),
("8C 9C 5C 3C TC", 19),
("JS QS 9H TS KH", 18),
("7C 7S KH 2H 7H", 17),
("3C KH 5D 5S KH", 16),
("QH 8H KD JH 8S", 15),
("2D 6D 9D TH 7D", 14),
)
def A_ ( ):
'''simple docstring'''
snake_case_, snake_case_ :Tuple = randrange(len(_lowercase ) ), randrange(len(_lowercase ) )
snake_case_ :Optional[Any] = ["""Loss""", """Tie""", """Win"""][(play >= oppo) + (play > oppo)]
snake_case_, snake_case_ :List[str] = SORTED_HANDS[play], SORTED_HANDS[oppo]
return hand, other, expected
def A_ ( _lowercase = 100 ):
'''simple docstring'''
return (generate_random_hand() for _ in range(_lowercase ))
@pytest.mark.parametrize("""hand, expected""", _lowercase )
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
assert PokerHand(_lowercase )._is_flush() == expected
@pytest.mark.parametrize("""hand, expected""", _lowercase )
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
assert PokerHand(_lowercase )._is_straight() == expected
@pytest.mark.parametrize("""hand, expected, card_values""", _lowercase )
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :Union[str, Any] = PokerHand(_lowercase )
assert player._is_five_high_straight() == expected
assert player._card_values == card_values
@pytest.mark.parametrize("""hand, expected""", _lowercase )
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
assert PokerHand(_lowercase )._is_same_kind() == expected
@pytest.mark.parametrize("""hand, expected""", _lowercase )
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
assert PokerHand(_lowercase )._hand_type == expected
@pytest.mark.parametrize("""hand, other, expected""", _lowercase )
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
assert PokerHand(_lowercase ).compare_with(PokerHand(_lowercase ) ) == expected
@pytest.mark.parametrize("""hand, other, expected""", generate_random_hands() )
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
assert PokerHand(_lowercase ).compare_with(PokerHand(_lowercase ) ) == expected
def A_ ( ):
'''simple docstring'''
snake_case_ :Any = [PokerHand(_lowercase ) for hand in SORTED_HANDS]
snake_case_ :Optional[Any] = poker_hands.copy()
shuffle(_lowercase )
snake_case_ :Optional[int] = chain(sorted(_lowercase ) )
for index, hand in enumerate(_lowercase ):
assert hand == poker_hands[index]
def A_ ( ):
'''simple docstring'''
snake_case_ :Any = [PokerHand("""2D AC 3H 4H 5S""" ), PokerHand("""2S 3H 4H 5S 6C""" )]
pokerhands.sort(reverse=_lowercase )
assert pokerhands[0].__str__() == "2S 3H 4H 5S 6C"
def A_ ( ):
'''simple docstring'''
snake_case_ :Any = PokerHand("""2C 4S AS 3D 5C""" )
snake_case_ :Optional[int] = True
snake_case_ :Optional[Any] = [5, 4, 3, 2, 14]
for _ in range(10 ):
assert pokerhand._is_five_high_straight() == expected
assert pokerhand._card_values == expected_card_values
def A_ ( ):
'''simple docstring'''
snake_case_ :Any = 0
snake_case_ :int = os.path.abspath(os.path.dirname(_lowercase ) )
snake_case_ :Optional[int] = os.path.join(_lowercase, """poker_hands.txt""" )
with open(_lowercase ) as file_hand:
for line in file_hand:
snake_case_ :Any = line[:14].strip()
snake_case_ :Union[str, Any] = line[15:].strip()
snake_case_, snake_case_ :Tuple = PokerHand(_lowercase ), PokerHand(_lowercase )
snake_case_ :List[str] = player.compare_with(_lowercase )
if output == "Win":
answer += 1
assert answer == 376
| 66 |
"""simple docstring"""
import math
class lowerCamelCase :
'''simple docstring'''
def lowerCAmelCase_ ( self: Tuple , snake_case: list[list[float]] , snake_case: list[int] ) -> int:
snake_case_ :Any = 0.0
snake_case_ :Tuple = 0.0
for i in range(len(snake_case ) ):
da += math.pow((sample[i] - weights[0][i]) , 2 )
da += math.pow((sample[i] - weights[1][i]) , 2 )
return 0 if da > da else 1
return 0
def lowerCAmelCase_ ( self: Optional[int] , snake_case: list[list[int | float]] , snake_case: list[int] , snake_case: int , snake_case: float ) -> list[list[int | float]]:
for i in range(len(snake_case ) ):
weights[j][i] += alpha * (sample[i] - weights[j][i])
return weights
def A_ ( ):
'''simple docstring'''
snake_case_ :Dict = [[1, 1, 0, 0], [0, 0, 0, 1], [1, 0, 0, 0], [0, 0, 1, 1]]
# weight initialization ( n, C )
snake_case_ :List[Any] = [[0.2, 0.6, 0.5, 0.9], [0.8, 0.4, 0.7, 0.3]]
# training
snake_case_ :Optional[Any] = SelfOrganizingMap()
snake_case_ :Dict = 3
snake_case_ :Dict = 0.5
for _ in range(_lowercase ):
for j in range(len(_lowercase ) ):
# training sample
snake_case_ :List[Any] = training_samples[j]
# Compute the winning vector
snake_case_ :Optional[int] = self_organizing_map.get_winner(_lowercase, _lowercase )
# Update the winning vector
snake_case_ :List[str] = self_organizing_map.update(_lowercase, _lowercase, _lowercase, _lowercase )
# classify test sample
snake_case_ :str = [0, 0, 0, 1]
snake_case_ :List[Any] = self_organizing_map.get_winner(_lowercase, _lowercase )
# results
print(f"""Clusters that the test sample belongs to : {winner}""" )
print(f"""Weights that have been trained : {weights}""" )
# running the main() function
if __name__ == "__main__":
main()
| 66 | 1 |
"""simple docstring"""
import copy
import os
from typing import Union
from ...configuration_utils import PretrainedConfig
from ...utils import logging
__a = logging.get_logger(__name__)
__a = {
"BAAI/AltCLIP": "https://huggingface.co/BAAI/AltCLIP/resolve/main/config.json",
# See all AltCLIP models at https://huggingface.co/models?filter=altclip
}
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
_A : Dict = """altclip_text_model"""
def __init__( self: Tuple , snake_case: List[Any]=250_002 , snake_case: Dict=1_024 , snake_case: Tuple=24 , snake_case: int=16 , snake_case: Any=4_096 , snake_case: Optional[int]="gelu" , snake_case: List[Any]=0.1 , snake_case: Union[str, Any]=0.1 , snake_case: List[str]=514 , snake_case: List[Any]=1 , snake_case: Any=0.0_2 , snake_case: Optional[int]=0.0_2 , snake_case: Union[str, Any]=1E-05 , snake_case: List[Any]=1 , snake_case: Tuple=0 , snake_case: int=2 , snake_case: Dict="absolute" , snake_case: Optional[int]=True , snake_case: Tuple=768 , **snake_case: Union[str, Any] , ) -> Union[str, Any]:
super().__init__(pad_token_id=snake_case , bos_token_id=snake_case , eos_token_id=snake_case , **snake_case )
snake_case_ :Dict = vocab_size
snake_case_ :str = hidden_size
snake_case_ :List[Any] = num_hidden_layers
snake_case_ :List[Any] = num_attention_heads
snake_case_ :Union[str, Any] = hidden_act
snake_case_ :Tuple = intermediate_size
snake_case_ :List[str] = hidden_dropout_prob
snake_case_ :int = attention_probs_dropout_prob
snake_case_ :int = max_position_embeddings
snake_case_ :Optional[int] = type_vocab_size
snake_case_ :Union[str, Any] = initializer_range
snake_case_ :str = initializer_factor
snake_case_ :List[str] = layer_norm_eps
snake_case_ :Tuple = position_embedding_type
snake_case_ :str = use_cache
snake_case_ :List[str] = project_dim
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
_A : Optional[Any] = """altclip_vision_model"""
def __init__( self: Dict , snake_case: Dict=768 , snake_case: List[Any]=3_072 , snake_case: int=512 , snake_case: List[Any]=12 , snake_case: List[str]=12 , snake_case: Optional[int]=3 , snake_case: int=224 , snake_case: Optional[Any]=32 , snake_case: Optional[int]="quick_gelu" , snake_case: Tuple=1E-5 , snake_case: Tuple=0.0 , snake_case: List[str]=0.0_2 , snake_case: Union[str, Any]=1.0 , **snake_case: str , ) -> Optional[Any]:
super().__init__(**snake_case )
snake_case_ :int = hidden_size
snake_case_ :Optional[int] = intermediate_size
snake_case_ :List[str] = projection_dim
snake_case_ :Any = num_hidden_layers
snake_case_ :Optional[Any] = num_attention_heads
snake_case_ :Union[str, Any] = num_channels
snake_case_ :int = patch_size
snake_case_ :List[Any] = image_size
snake_case_ :int = initializer_range
snake_case_ :Optional[Any] = initializer_factor
snake_case_ :str = attention_dropout
snake_case_ :List[Any] = layer_norm_eps
snake_case_ :List[str] = hidden_act
@classmethod
def lowerCAmelCase_ ( cls: List[str] , snake_case: Union[str, os.PathLike] , **snake_case: List[Any] ) -> "PretrainedConfig":
cls._set_token_in_kwargs(snake_case )
snake_case_, snake_case_ :List[str] = cls.get_config_dict(snake_case , **snake_case )
# get the vision config dict if we are loading from AltCLIPConfig
if config_dict.get("""model_type""" ) == "altclip":
snake_case_ :Optional[Any] = config_dict["""vision_config"""]
if "model_type" in config_dict and hasattr(cls , """model_type""" ) and config_dict["model_type"] != cls.model_type:
logger.warning(
f"""You are using a model of type {config_dict['model_type']} to instantiate a model of type """
f"""{cls.model_type}. This is not supported for all configurations of models and can yield errors.""" )
return cls.from_dict(snake_case , **snake_case )
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
_A : str = """altclip"""
_A : Any = True
def __init__( self: int , snake_case: Any=None , snake_case: Optional[Any]=None , snake_case: str=768 , snake_case: int=2.6_5_9_2 , **snake_case: Union[str, Any] ) -> str:
# If `_config_dict` exist, we use them for the backward compatibility.
# We pop out these 2 attributes before calling `super().__init__` to avoid them being saved (which causes a lot
# of confusion!).
snake_case_ :str = kwargs.pop("""text_config_dict""" , snake_case )
snake_case_ :List[str] = kwargs.pop("""vision_config_dict""" , snake_case )
super().__init__(**snake_case )
# Instead of simply assigning `[text|vision]_config_dict` to `[text|vision]_config`, we use the values in
# `[text|vision]_config_dict` to update the values in `[text|vision]_config`. The values should be same in most
# cases, but we don't want to break anything regarding `_config_dict` that existed before commit `8827e1b2`.
if text_config_dict is not None:
if text_config is None:
snake_case_ :Tuple = {}
# This is the complete result when using `text_config_dict`.
snake_case_ :str = AltCLIPTextConfig(**snake_case ).to_dict()
# Give a warning if the values exist in both `_text_config_dict` and `text_config` but being different.
for key, value in _text_config_dict.items():
if key in text_config and value != text_config[key] and key not in ["transformers_version"]:
# If specified in `text_config_dict`
if key in text_config_dict:
snake_case_ :List[str] = (
f"""`{key}` is found in both `text_config_dict` and `text_config` but with different values. """
f"""The value `text_config_dict[\"{key}\"]` will be used instead."""
)
# If inferred from default argument values (just to be super careful)
else:
snake_case_ :str = (
f"""`text_config_dict` is provided which will be used to initialize `AltCLIPTextConfig`. The """
f"""value `text_config[\"{key}\"]` will be overriden."""
)
logger.warning(snake_case )
# Update all values in `text_config` with the ones in `_text_config_dict`.
text_config.update(_text_config_dict )
if vision_config_dict is not None:
if vision_config is None:
snake_case_ :Tuple = {}
# This is the complete result when using `vision_config_dict`.
snake_case_ :Dict = AltCLIPVisionConfig(**snake_case ).to_dict()
# convert keys to string instead of integer
if "id2label" in _vision_config_dict:
snake_case_ :Union[str, Any] = {
str(snake_case ): value for key, value in _vision_config_dict["""id2label"""].items()
}
# Give a warning if the values exist in both `_vision_config_dict` and `vision_config` but being different.
for key, value in _vision_config_dict.items():
if key in vision_config and value != vision_config[key] and key not in ["transformers_version"]:
# If specified in `vision_config_dict`
if key in vision_config_dict:
snake_case_ :Dict = (
f"""`{key}` is found in both `vision_config_dict` and `vision_config` but with different """
f"""values. The value `vision_config_dict[\"{key}\"]` will be used instead."""
)
# If inferred from default argument values (just to be super careful)
else:
snake_case_ :Dict = (
f"""`vision_config_dict` is provided which will be used to initialize `AltCLIPVisionConfig`. """
f"""The value `vision_config[\"{key}\"]` will be overriden."""
)
logger.warning(snake_case )
# Update all values in `vision_config` with the ones in `_vision_config_dict`.
vision_config.update(_vision_config_dict )
if text_config is None:
snake_case_ :Optional[Any] = {}
logger.info("""`text_config` is `None`. Initializing the `AltCLIPTextConfig` with default values.""" )
if vision_config is None:
snake_case_ :List[str] = {}
logger.info("""`vision_config` is `None`. initializing the `AltCLIPVisionConfig` with default values.""" )
snake_case_ :Tuple = AltCLIPTextConfig(**snake_case )
snake_case_ :str = AltCLIPVisionConfig(**snake_case )
snake_case_ :int = projection_dim
snake_case_ :Union[str, Any] = logit_scale_init_value
snake_case_ :Dict = 1.0
@classmethod
def lowerCAmelCase_ ( cls: str , snake_case: AltCLIPTextConfig , snake_case: AltCLIPVisionConfig , **snake_case: int ) -> Union[str, Any]:
return cls(text_config=text_config.to_dict() , vision_config=vision_config.to_dict() , **snake_case )
def lowerCAmelCase_ ( self: int ) -> Tuple:
snake_case_ :str = copy.deepcopy(self.__dict__ )
snake_case_ :Any = self.text_config.to_dict()
snake_case_ :Optional[int] = self.vision_config.to_dict()
snake_case_ :List[Any] = self.__class__.model_type
return output
| 66 |
"""simple docstring"""
import collections
import inspect
import unittest
from transformers import SwinvaConfig
from transformers.testing_utils import require_torch, require_vision, slow, torch_device
from transformers.utils import cached_property, is_torch_available, is_vision_available
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, _config_zero_init, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from torch import nn
from transformers import SwinvaForImageClassification, SwinvaForMaskedImageModeling, SwinvaModel
from transformers.models.swinva.modeling_swinva import SWINV2_PRETRAINED_MODEL_ARCHIVE_LIST
if is_vision_available():
from PIL import Image
from transformers import AutoImageProcessor
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Optional[int] , snake_case: Any , snake_case: Optional[Any]=13 , snake_case: Tuple=32 , snake_case: Optional[int]=2 , snake_case: Tuple=3 , snake_case: Tuple=16 , snake_case: Optional[Any]=[1, 2, 1] , snake_case: Optional[int]=[2, 2, 4] , snake_case: Optional[int]=2 , snake_case: int=2.0 , snake_case: Union[str, Any]=True , snake_case: List[str]=0.0 , snake_case: List[Any]=0.0 , snake_case: Optional[Any]=0.1 , snake_case: List[Any]="gelu" , snake_case: Optional[int]=False , snake_case: Union[str, Any]=True , snake_case: Union[str, Any]=0.0_2 , snake_case: Optional[int]=1E-5 , snake_case: Optional[Any]=True , snake_case: List[Any]=None , snake_case: List[Any]=True , snake_case: Optional[Any]=10 , snake_case: str=8 , ) -> Tuple:
snake_case_ :Dict = parent
snake_case_ :Any = batch_size
snake_case_ :List[Any] = image_size
snake_case_ :List[Any] = patch_size
snake_case_ :int = num_channels
snake_case_ :Tuple = embed_dim
snake_case_ :str = depths
snake_case_ :str = num_heads
snake_case_ :Optional[int] = window_size
snake_case_ :Tuple = mlp_ratio
snake_case_ :Any = qkv_bias
snake_case_ :List[Any] = hidden_dropout_prob
snake_case_ :Optional[Any] = attention_probs_dropout_prob
snake_case_ :Union[str, Any] = drop_path_rate
snake_case_ :Any = hidden_act
snake_case_ :Optional[Any] = use_absolute_embeddings
snake_case_ :Union[str, Any] = patch_norm
snake_case_ :Dict = layer_norm_eps
snake_case_ :str = initializer_range
snake_case_ :Tuple = is_training
snake_case_ :Tuple = scope
snake_case_ :Union[str, Any] = use_labels
snake_case_ :Optional[Any] = type_sequence_label_size
snake_case_ :Dict = encoder_stride
def lowerCAmelCase_ ( self: int ) -> int:
snake_case_ :List[str] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case_ :Any = None
if self.use_labels:
snake_case_ :str = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ :int = self.get_config()
return config, pixel_values, labels
def lowerCAmelCase_ ( self: str ) -> Union[str, Any]:
return SwinvaConfig(
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 , )
def lowerCAmelCase_ ( self: str , snake_case: Optional[int] , snake_case: Dict , snake_case: str ) -> List[Any]:
snake_case_ :Union[str, Any] = SwinvaModel(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Optional[int] = model(snake_case )
snake_case_ :Optional[Any] = ((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 lowerCAmelCase_ ( self: int , snake_case: List[str] , snake_case: Tuple , snake_case: int ) -> Any:
snake_case_ :Dict = SwinvaForMaskedImageModeling(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Tuple = model(snake_case )
self.parent.assertEqual(
result.logits.shape , (self.batch_size, self.num_channels, self.image_size, self.image_size) )
# test greyscale images
snake_case_ :List[Any] = 1
snake_case_ :int = SwinvaForMaskedImageModeling(snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Union[str, Any] = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] )
snake_case_ :int = model(snake_case )
self.parent.assertEqual(result.logits.shape , (self.batch_size, 1, self.image_size, self.image_size) )
def lowerCAmelCase_ ( self: List[Any] , snake_case: Any , snake_case: List[str] , snake_case: Union[str, Any] ) -> Tuple:
snake_case_ :int = self.type_sequence_label_size
snake_case_ :List[Any] = SwinvaForImageClassification(snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Dict = model(snake_case , labels=snake_case )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
def lowerCAmelCase_ ( self: int ) -> str:
snake_case_ :Any = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_ :List[str] = config_and_inputs
snake_case_ :List[Any] = {"""pixel_values""": pixel_values}
return config, inputs_dict
@require_torch
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : Optional[Any] = (
(SwinvaModel, SwinvaForImageClassification, SwinvaForMaskedImageModeling) if is_torch_available() else ()
)
_A : Any = (
{"""feature-extraction""": SwinvaModel, """image-classification""": SwinvaForImageClassification}
if is_torch_available()
else {}
)
_A : List[Any] = False
_A : List[str] = False
_A : Tuple = False
_A : List[str] = False
def lowerCAmelCase_ ( self: Dict ) -> List[Any]:
snake_case_ :Optional[int] = SwinvaModelTester(self )
snake_case_ :List[str] = ConfigTester(self , config_class=snake_case , embed_dim=37 )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> List[Any]:
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 lowerCAmelCase_ ( self: Union[str, Any] ) -> Tuple:
snake_case_ :List[str] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case )
@unittest.skip(reason="""Got `CUDA error: misaligned address` with PyTorch 2.0.0.""" )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> str:
pass
@unittest.skip(reason="""Swinv2 does not use inputs_embeds""" )
def lowerCAmelCase_ ( self: int ) -> Dict:
pass
def lowerCAmelCase_ ( self: List[str] ) -> Union[str, Any]:
snake_case_, snake_case_ :List[str] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :Optional[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 lowerCAmelCase_ ( self: Dict ) -> Optional[int]:
snake_case_, snake_case_ :Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :Optional[int] = model_class(snake_case )
snake_case_ :List[Any] = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case_ :int = [*signature.parameters.keys()]
snake_case_ :List[Any] = ["""pixel_values"""]
self.assertListEqual(arg_names[:1] , snake_case )
def lowerCAmelCase_ ( self: List[str] ) -> Optional[Any]:
snake_case_, snake_case_ :List[str] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :List[str] = True
for model_class in self.all_model_classes:
snake_case_ :List[Any] = True
snake_case_ :Any = False
snake_case_ :Optional[int] = True
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_ :str = outputs.attentions
snake_case_ :Dict = len(self.model_tester.depths )
self.assertEqual(len(snake_case ) , snake_case )
# check that output_attentions also work using config
del inputs_dict["output_attentions"]
snake_case_ :Union[str, Any] = True
snake_case_ :Tuple = config.window_size**2
snake_case_ :Any = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :Union[str, Any] = model(**self._prepare_for_class(snake_case , snake_case ) )
snake_case_ :int = outputs.attentions
self.assertEqual(len(snake_case ) , snake_case )
self.assertListEqual(
list(attentions[0].shape[-3:] ) , [self.model_tester.num_heads[0], window_size_squared, window_size_squared] , )
snake_case_ :Any = len(snake_case )
# Check attention is always last and order is fine
snake_case_ :int = True
snake_case_ :Dict = True
snake_case_ :Optional[int] = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :Dict = model(**self._prepare_for_class(snake_case , snake_case ) )
if hasattr(self.model_tester , """num_hidden_states_types""" ):
snake_case_ :Any = self.model_tester.num_hidden_states_types
else:
# also another +1 for reshaped_hidden_states
snake_case_ :int = 2
self.assertEqual(out_len + added_hidden_states , len(snake_case ) )
snake_case_ :str = outputs.attentions
self.assertEqual(len(snake_case ) , snake_case )
self.assertListEqual(
list(self_attentions[0].shape[-3:] ) , [self.model_tester.num_heads[0], window_size_squared, window_size_squared] , )
def lowerCAmelCase_ ( self: int , snake_case: Dict , snake_case: Dict , snake_case: Optional[Any] , snake_case: Dict ) -> List[str]:
snake_case_ :Dict = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :Optional[int] = model(**self._prepare_for_class(snake_case , snake_case ) )
snake_case_ :str = outputs.hidden_states
snake_case_ :List[Any] = getattr(
self.model_tester , """expected_num_hidden_layers""" , len(self.model_tester.depths ) + 1 )
self.assertEqual(len(snake_case ) , snake_case )
# Swinv2 has a different seq_length
snake_case_ :List[Any] = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :Optional[int] = (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] , )
snake_case_ :str = outputs.reshaped_hidden_states
self.assertEqual(len(snake_case ) , snake_case )
snake_case_, snake_case_, snake_case_, snake_case_ :Any = reshaped_hidden_states[0].shape
snake_case_ :int = (
reshaped_hidden_states[0].view(snake_case , snake_case , height * width ).permute(0 , 2 , 1 )
)
self.assertListEqual(
list(reshaped_hidden_states.shape[-2:] ) , [num_patches, self.model_tester.embed_dim] , )
def lowerCAmelCase_ ( self: Any ) -> Any:
snake_case_, snake_case_ :List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Union[str, Any] = (
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_ :Union[str, Any] = 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_ :List[str] = True
self.check_hidden_states_output(snake_case , snake_case , snake_case , snake_case )
def lowerCAmelCase_ ( self: Tuple ) -> Any:
snake_case_, snake_case_ :Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Optional[int] = 3
snake_case_ :Union[str, Any] = (
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_ :str = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :Any = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
snake_case_ :int = 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_ :Tuple = True
self.check_hidden_states_output(snake_case , snake_case , snake_case , (padded_height, padded_width) )
def lowerCAmelCase_ ( self: Any ) -> Tuple:
snake_case_ :int = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_masked_image_modeling(*snake_case )
def lowerCAmelCase_ ( self: Optional[int] ) -> Dict:
snake_case_ :Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_classification(*snake_case )
@slow
def lowerCAmelCase_ ( self: List[Any] ) -> Dict:
for model_name in SWINV2_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
snake_case_ :List[str] = SwinvaModel.from_pretrained(snake_case )
self.assertIsNotNone(snake_case )
def lowerCAmelCase_ ( self: Optional[int] ) -> List[Any]:
snake_case_, snake_case_ :str = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Optional[int] = _config_zero_init(snake_case )
for model_class in self.all_model_classes:
snake_case_ :Tuple = model_class(config=snake_case )
for name, param in model.named_parameters():
if "embeddings" not in name and "logit_scale" not in name and param.requires_grad:
self.assertIn(
((param.data.mean() * 1E9).round() / 1E9).item() , [0.0, 1.0] , msg=f"""Parameter {name} of model {model_class} seems not properly initialized""" , )
@require_vision
@require_torch
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
@cached_property
def lowerCAmelCase_ ( self: Optional[int] ) -> List[Any]:
return (
AutoImageProcessor.from_pretrained("""microsoft/swinv2-tiny-patch4-window8-256""" )
if is_vision_available()
else None
)
@slow
def lowerCAmelCase_ ( self: List[str] ) -> List[str]:
snake_case_ :Tuple = SwinvaForImageClassification.from_pretrained("""microsoft/swinv2-tiny-patch4-window8-256""" ).to(
snake_case )
snake_case_ :str = self.default_image_processor
snake_case_ :List[Any] = Image.open("""./tests/fixtures/tests_samples/COCO/000000039769.png""" )
snake_case_ :str = image_processor(images=snake_case , return_tensors="""pt""" ).to(snake_case )
# forward pass
with torch.no_grad():
snake_case_ :Tuple = model(**snake_case )
# verify the logits
snake_case_ :Dict = torch.Size((1, 1_000) )
self.assertEqual(outputs.logits.shape , snake_case )
snake_case_ :int = torch.tensor([-0.3_9_4_7, -0.4_3_0_6, 0.0_0_2_6] ).to(snake_case )
self.assertTrue(torch.allclose(outputs.logits[0, :3] , snake_case , atol=1E-4 ) )
| 66 | 1 |
"""simple docstring"""
import numpy as np
from scipy.spatial.distance import cdist
from sklearn.metrics import fa_score
import datasets
__a = "\\n @inproceedings{kakwani2020indicnlpsuite,\n title={{IndicNLPSuite: Monolingual Corpora, Evaluation Benchmarks and Pre-trained Multilingual Language Models for Indian Languages}},\n author={Divyanshu Kakwani and Anoop Kunchukuttan and Satish Golla and Gokul N.C. and Avik Bhattacharyya and Mitesh M. Khapra and Pratyush Kumar},\n year={2020},\n booktitle={Findings of EMNLP},\n}\n"
__a = "\\n IndicGLUE is a natural language understanding benchmark for Indian languages. It contains a wide\n variety of tasks and covers 11 major Indian languages - as, bn, gu, hi, kn, ml, mr, or, pa, ta, te.\n"
__a = "\nCompute IndicGLUE evaluation metric associated to each IndicGLUE dataset.\nArgs:\n predictions: list of predictions to score (as int64),\n except for 'cvit-mkb-clsr' where each prediction is a vector (of float32).\n references: list of ground truth labels corresponding to the predictions (as int64),\n except for 'cvit-mkb-clsr' where each reference is a vector (of float32).\nReturns: depending on the IndicGLUE subset, one or several of:\n \"accuracy\": Accuracy\n \"f1\": F1 score\n \"precision\": Precision@10\nExamples:\n\n >>> indic_glue_metric = datasets.load_metric('indic_glue', 'wnli') # 'wnli' or any of [\"copa\", \"sna\", \"csqa\", \"wstp\", \"inltkh\", \"bbca\", \"iitp-mr\", \"iitp-pr\", \"actsa-sc\", \"md\"]\n >>> references = [0, 1]\n >>> predictions = [0, 1]\n >>> results = indic_glue_metric.compute(predictions=predictions, references=references)\n >>> print(results)\n {'accuracy': 1.0}\n\n >>> indic_glue_metric = datasets.load_metric('indic_glue', 'wiki-ner')\n >>> references = [0, 1]\n >>> predictions = [0, 1]\n >>> results = indic_glue_metric.compute(predictions=predictions, references=references)\n >>> print(results)\n {'accuracy': 1.0, 'f1': 1.0}\n\n >>> indic_glue_metric = datasets.load_metric('indic_glue', 'cvit-mkb-clsr')\n >>> references = [[0.5, 0.5, 0.5], [0.1, 0.2, 0.3]]\n >>> predictions = [[0.5, 0.5, 0.5], [0.1, 0.2, 0.3]]\n >>> results = indic_glue_metric.compute(predictions=predictions, references=references)\n >>> print(results)\n {'precision@10': 1.0}\n\n"
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
return float((preds == labels).mean() )
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :Any = simple_accuracy(_lowercase, _lowercase )
snake_case_ :List[str] = float(fa_score(y_true=_lowercase, y_pred=_lowercase ) )
return {
"accuracy": acc,
"f1": fa,
}
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :Union[str, Any] = np.array(_lowercase )
snake_case_ :Tuple = np.array(_lowercase )
snake_case_ :Dict = en_sentvecs.shape[0]
# mean centering
snake_case_ :int = en_sentvecs - np.mean(_lowercase, axis=0 )
snake_case_ :Dict = in_sentvecs - np.mean(_lowercase, axis=0 )
snake_case_ :int = cdist(_lowercase, _lowercase, """cosine""" )
snake_case_ :Tuple = np.array(range(_lowercase ) )
snake_case_ :Optional[Any] = sim.argsort(axis=1 )[:, :10]
snake_case_ :List[str] = np.any(preds == actual[:, None], axis=1 )
return float(matches.mean() )
@datasets.utils.file_utils.add_start_docstrings(_DESCRIPTION , _KWARGS_DESCRIPTION )
class lowerCamelCase ( datasets.Metric ):
'''simple docstring'''
def lowerCAmelCase_ ( self: List[str] ) -> Union[str, Any]:
if self.config_name not in [
"wnli",
"copa",
"sna",
"csqa",
"wstp",
"inltkh",
"bbca",
"cvit-mkb-clsr",
"iitp-mr",
"iitp-pr",
"actsa-sc",
"md",
"wiki-ner",
]:
raise KeyError(
"""You should supply a configuration name selected in """
"""[\"wnli\", \"copa\", \"sna\", \"csqa\", \"wstp\", \"inltkh\", \"bbca\", """
"""\"cvit-mkb-clsr\", \"iitp-mr\", \"iitp-pr\", \"actsa-sc\", \"md\", """
"""\"wiki-ner\"]""" )
return datasets.MetricInfo(
description=_DESCRIPTION , citation=_CITATION , inputs_description=_KWARGS_DESCRIPTION , features=datasets.Features(
{
"""predictions""": datasets.Value("""int64""" )
if self.config_name != """cvit-mkb-clsr"""
else datasets.Sequence(datasets.Value("""float32""" ) ),
"""references""": datasets.Value("""int64""" )
if self.config_name != """cvit-mkb-clsr"""
else datasets.Sequence(datasets.Value("""float32""" ) ),
} ) , codebase_urls=[] , reference_urls=[] , format="""numpy""" if self.config_name != """cvit-mkb-clsr""" else None , )
def lowerCAmelCase_ ( self: int , snake_case: Tuple , snake_case: Optional[int] ) -> Tuple:
if self.config_name == "cvit-mkb-clsr":
return {"precision@10": precision_at_aa(snake_case , snake_case )}
elif self.config_name in ["wiki-ner"]:
return acc_and_fa(snake_case , snake_case )
elif self.config_name in [
"wnli",
"copa",
"sna",
"csqa",
"wstp",
"inltkh",
"bbca",
"iitp-mr",
"iitp-pr",
"actsa-sc",
"md",
]:
return {"accuracy": simple_accuracy(snake_case , snake_case )}
else:
raise KeyError(
"""You should supply a configuration name selected in """
"""[\"wnli\", \"copa\", \"sna\", \"csqa\", \"wstp\", \"inltkh\", \"bbca\", """
"""\"cvit-mkb-clsr\", \"iitp-mr\", \"iitp-pr\", \"actsa-sc\", \"md\", """
"""\"wiki-ner\"]""" )
| 66 |
"""simple docstring"""
import re
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Optional[int] = re.compile(
r"""^(?:0|94|\+94|0{2}94)""" r"""7(0|1|2|4|5|6|7|8)""" r"""(-| |)""" r"""\d{7}$""" )
return bool(re.search(_lowercase, _lowercase ) )
if __name__ == "__main__":
__a = "0094702343221"
print(is_sri_lankan_phone_number(phone))
| 66 | 1 |
"""simple docstring"""
from math import factorial
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Optional[int] , snake_case: Dict , snake_case: int ) -> Tuple:
snake_case_ :List[Any] = real
if isinstance(snake_case , snake_case ):
snake_case_ :Tuple = [1] * rank
else:
snake_case_ :Optional[Any] = rank
def __repr__( self: List[str] ) -> Tuple:
return (
f"""{self.real}+"""
f"""{'+'.join(str(snake_case )+'E'+str(n+1 )for n,dual in enumerate(self.duals ) )}"""
)
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
snake_case_ :Any = self.duals.copy()
while cur[-1] == 0:
cur.pop(-1 )
return Dual(self.real , snake_case )
def __add__( self: Optional[int] , snake_case: Dict ) -> List[str]:
if not isinstance(snake_case , snake_case ):
return Dual(self.real + other , self.duals )
snake_case_ :List[Any] = self.duals.copy()
snake_case_ :Tuple = other.duals.copy()
if len(snake_case ) > len(snake_case ):
o_dual.extend([1] * (len(snake_case ) - len(snake_case )) )
elif len(snake_case ) < len(snake_case ):
s_dual.extend([1] * (len(snake_case ) - len(snake_case )) )
snake_case_ :Dict = []
for i in range(len(snake_case ) ):
new_duals.append(s_dual[i] + o_dual[i] )
return Dual(self.real + other.real , snake_case )
_A : str = __add__
def __sub__( self: Tuple , snake_case: Union[str, Any] ) -> Tuple:
return self + other * -1
def __mul__( self: str , snake_case: Tuple ) -> Optional[Any]:
if not isinstance(snake_case , snake_case ):
snake_case_ :Dict = []
for i in self.duals:
new_duals.append(i * other )
return Dual(self.real * other , snake_case )
snake_case_ :int = [0] * (len(self.duals ) + len(other.duals ) + 1)
for i, item in enumerate(self.duals ):
for j, jtem in enumerate(other.duals ):
new_duals[i + j + 1] += item * jtem
for k in range(len(self.duals ) ):
new_duals[k] += self.duals[k] * other.real
for index in range(len(other.duals ) ):
new_duals[index] += other.duals[index] * self.real
return Dual(self.real * other.real , snake_case )
_A : int = __mul__
def __truediv__( self: List[str] , snake_case: List[str] ) -> List[str]:
if not isinstance(snake_case , snake_case ):
snake_case_ :Optional[Any] = []
for i in self.duals:
new_duals.append(i / other )
return Dual(self.real / other , snake_case )
raise ValueError
def __floordiv__( self: int , snake_case: List[Any] ) -> Any:
if not isinstance(snake_case , snake_case ):
snake_case_ :Optional[int] = []
for i in self.duals:
new_duals.append(i // other )
return Dual(self.real // other , snake_case )
raise ValueError
def __pow__( self: Optional[Any] , snake_case: Optional[int] ) -> List[Any]:
if n < 0 or isinstance(snake_case , snake_case ):
raise ValueError("""power must be a positive integer""" )
if n == 0:
return 1
if n == 1:
return self
snake_case_ :str = self
for _ in range(n - 1 ):
x *= self
return x
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
if not callable(_lowercase ):
raise ValueError("""differentiate() requires a function as input for func""" )
if not isinstance(_lowercase, (float, int) ):
raise ValueError("""differentiate() requires a float as input for position""" )
if not isinstance(_lowercase, _lowercase ):
raise ValueError("""differentiate() requires an int as input for order""" )
snake_case_ :Optional[Any] = Dual(_lowercase, 1 )
snake_case_ :List[Any] = func(_lowercase )
if order == 0:
return result.real
return result.duals[order - 1] * factorial(_lowercase )
if __name__ == "__main__":
import doctest
doctest.testmod()
def A_ ( _lowercase ):
'''simple docstring'''
return y**2 * y**4
print(differentiate(f, 9, 2))
| 66 |
"""simple docstring"""
import argparse
import json
import os
import pickle
import shutil
import numpy as np
import torch
from distiller import Distiller
from lm_seqs_dataset import LmSeqsDataset
from transformers import (
BertConfig,
BertForMaskedLM,
BertTokenizer,
DistilBertConfig,
DistilBertForMaskedLM,
DistilBertTokenizer,
GPTaConfig,
GPTaLMHeadModel,
GPTaTokenizer,
RobertaConfig,
RobertaForMaskedLM,
RobertaTokenizer,
)
from utils import git_log, init_gpu_params, logger, set_seed
__a = {
"distilbert": (DistilBertConfig, DistilBertForMaskedLM, DistilBertTokenizer),
"roberta": (RobertaConfig, RobertaForMaskedLM, RobertaTokenizer),
"bert": (BertConfig, BertForMaskedLM, BertTokenizer),
"gpt2": (GPTaConfig, GPTaLMHeadModel, GPTaTokenizer),
}
def A_ ( _lowercase ):
'''simple docstring'''
assert (args.mlm and args.alpha_mlm > 0.0) or (not args.mlm and args.alpha_mlm == 0.0)
assert (args.alpha_mlm > 0.0 and args.alpha_clm == 0.0) or (args.alpha_mlm == 0.0 and args.alpha_clm > 0.0)
if args.mlm:
assert os.path.isfile(args.token_counts )
assert (args.student_type in ["roberta", "distilbert"]) and (args.teacher_type in ["roberta", "bert"])
else:
assert (args.student_type in ["gpt2"]) and (args.teacher_type in ["gpt2"])
assert args.teacher_type == args.student_type or (
args.student_type == "distilbert" and args.teacher_type == "bert"
)
assert os.path.isfile(args.student_config )
if args.student_pretrained_weights is not None:
assert os.path.isfile(args.student_pretrained_weights )
if args.freeze_token_type_embds:
assert args.student_type in ["roberta"]
assert args.alpha_ce >= 0.0
assert args.alpha_mlm >= 0.0
assert args.alpha_clm >= 0.0
assert args.alpha_mse >= 0.0
assert args.alpha_cos >= 0.0
assert args.alpha_ce + args.alpha_mlm + args.alpha_clm + args.alpha_mse + args.alpha_cos > 0.0
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
if args.student_type == "roberta":
snake_case_ :Tuple = False
elif args.student_type == "gpt2":
snake_case_ :Union[str, Any] = False
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
if args.student_type == "roberta":
snake_case_ :List[str] = False
def A_ ( ):
'''simple docstring'''
snake_case_ :Union[str, Any] = argparse.ArgumentParser(description="""Training""" )
parser.add_argument("""--force""", action="""store_true""", help="""Overwrite dump_path if it already exists.""" )
parser.add_argument(
"""--dump_path""", type=_lowercase, required=_lowercase, help="""The output directory (log, checkpoints, parameters, etc.)""" )
parser.add_argument(
"""--data_file""", type=_lowercase, required=_lowercase, help="""The binarized file (tokenized + tokens_to_ids) and grouped by sequence.""", )
parser.add_argument(
"""--student_type""", type=_lowercase, choices=["""distilbert""", """roberta""", """gpt2"""], required=_lowercase, help="""The student type (DistilBERT, RoBERTa).""", )
parser.add_argument("""--student_config""", type=_lowercase, required=_lowercase, help="""Path to the student configuration.""" )
parser.add_argument(
"""--student_pretrained_weights""", default=_lowercase, type=_lowercase, help="""Load student initialization checkpoint.""" )
parser.add_argument(
"""--teacher_type""", choices=["""bert""", """roberta""", """gpt2"""], required=_lowercase, help="""Teacher type (BERT, RoBERTa).""" )
parser.add_argument("""--teacher_name""", type=_lowercase, required=_lowercase, help="""The teacher model.""" )
parser.add_argument("""--temperature""", default=2.0, type=_lowercase, help="""Temperature for the softmax temperature.""" )
parser.add_argument(
"""--alpha_ce""", default=0.5, type=_lowercase, help="""Linear weight for the distillation loss. Must be >=0.""" )
parser.add_argument(
"""--alpha_mlm""", default=0.0, type=_lowercase, help="""Linear weight for the MLM loss. Must be >=0. Should be used in conjunction with `mlm` flag.""", )
parser.add_argument("""--alpha_clm""", default=0.5, type=_lowercase, help="""Linear weight for the CLM loss. Must be >=0.""" )
parser.add_argument("""--alpha_mse""", default=0.0, type=_lowercase, help="""Linear weight of the MSE loss. Must be >=0.""" )
parser.add_argument(
"""--alpha_cos""", default=0.0, type=_lowercase, help="""Linear weight of the cosine embedding loss. Must be >=0.""" )
parser.add_argument(
"""--mlm""", action="""store_true""", help="""The LM step: MLM or CLM. If `mlm` is True, the MLM is used over CLM.""" )
parser.add_argument(
"""--mlm_mask_prop""", default=0.15, type=_lowercase, help="""Proportion of tokens for which we need to make a prediction.""", )
parser.add_argument("""--word_mask""", default=0.8, type=_lowercase, help="""Proportion of tokens to mask out.""" )
parser.add_argument("""--word_keep""", default=0.1, type=_lowercase, help="""Proportion of tokens to keep.""" )
parser.add_argument("""--word_rand""", default=0.1, type=_lowercase, help="""Proportion of tokens to randomly replace.""" )
parser.add_argument(
"""--mlm_smoothing""", default=0.7, type=_lowercase, help="""Smoothing parameter to emphasize more rare tokens (see XLM, similar to word2vec).""", )
parser.add_argument("""--token_counts""", type=_lowercase, help="""The token counts in the data_file for MLM.""" )
parser.add_argument(
"""--restrict_ce_to_mask""", action="""store_true""", help="""If true, compute the distillation loss only the [MLM] prediction distribution.""", )
parser.add_argument(
"""--freeze_pos_embs""", action="""store_true""", help="""Freeze positional embeddings during distillation. For student_type in ['roberta', 'gpt2'] only.""", )
parser.add_argument(
"""--freeze_token_type_embds""", action="""store_true""", help="""Freeze token type embeddings during distillation if existent. For student_type in ['roberta'] only.""", )
parser.add_argument("""--n_epoch""", type=_lowercase, default=3, help="""Number of pass on the whole dataset.""" )
parser.add_argument("""--batch_size""", type=_lowercase, default=5, help="""Batch size (for each process).""" )
parser.add_argument(
"""--group_by_size""", action="""store_false""", help="""If true, group sequences that have similar length into the same batch. Default is true.""", )
parser.add_argument(
"""--gradient_accumulation_steps""", type=_lowercase, default=50, help="""Gradient accumulation for larger training batches.""", )
parser.add_argument("""--warmup_prop""", default=0.05, type=_lowercase, help="""Linear warmup proportion.""" )
parser.add_argument("""--weight_decay""", default=0.0, type=_lowercase, help="""Weight decay if we apply some.""" )
parser.add_argument("""--learning_rate""", default=5e-4, type=_lowercase, help="""The initial learning rate for Adam.""" )
parser.add_argument("""--adam_epsilon""", default=1e-6, type=_lowercase, help="""Epsilon for Adam optimizer.""" )
parser.add_argument("""--max_grad_norm""", default=5.0, type=_lowercase, help="""Max gradient norm.""" )
parser.add_argument("""--initializer_range""", default=0.02, type=_lowercase, help="""Random initialization range.""" )
parser.add_argument(
"""--fp16""", action="""store_true""", help="""Whether to use 16-bit (mixed) precision (through NVIDIA apex) instead of 32-bit""", )
parser.add_argument(
"""--fp16_opt_level""", type=_lowercase, default="""O1""", help=(
"""For fp16: Apex AMP optimization level selected in ['O0', 'O1', 'O2', and 'O3']."""
"""See details at https://nvidia.github.io/apex/amp.html"""
), )
parser.add_argument("""--n_gpu""", type=_lowercase, default=1, help="""Number of GPUs in the node.""" )
parser.add_argument("""--local_rank""", type=_lowercase, default=-1, help="""Distributed training - Local rank""" )
parser.add_argument("""--seed""", type=_lowercase, default=56, help="""Random seed""" )
parser.add_argument("""--log_interval""", type=_lowercase, default=500, help="""Tensorboard logging interval.""" )
parser.add_argument("""--checkpoint_interval""", type=_lowercase, default=4000, help="""Checkpoint interval.""" )
snake_case_ :Tuple = parser.parse_args()
sanity_checks(_lowercase )
# ARGS #
init_gpu_params(_lowercase )
set_seed(_lowercase )
if args.is_master:
if os.path.exists(args.dump_path ):
if not args.force:
raise ValueError(
f"""Serialization dir {args.dump_path} already exists, but you have not precised wheter to overwrite"""
""" itUse `--force` if you want to overwrite it""" )
else:
shutil.rmtree(args.dump_path )
if not os.path.exists(args.dump_path ):
os.makedirs(args.dump_path )
logger.info(f"""Experiment will be dumped and logged in {args.dump_path}""" )
# SAVE PARAMS #
logger.info(f"""Param: {args}""" )
with open(os.path.join(args.dump_path, """parameters.json""" ), """w""" ) as f:
json.dump(vars(_lowercase ), _lowercase, indent=4 )
git_log(args.dump_path )
snake_case_, snake_case_, snake_case_ :Any = MODEL_CLASSES[args.student_type]
snake_case_, snake_case_, snake_case_ :int = MODEL_CLASSES[args.teacher_type]
# TOKENIZER #
snake_case_ :Any = teacher_tokenizer_class.from_pretrained(args.teacher_name )
snake_case_ :Optional[Any] = {}
for tok_name, tok_symbol in tokenizer.special_tokens_map.items():
snake_case_ :Union[str, Any] = tokenizer.all_special_tokens.index(_lowercase )
snake_case_ :Union[str, Any] = tokenizer.all_special_ids[idx]
logger.info(f"""Special tokens {special_tok_ids}""" )
snake_case_ :str = special_tok_ids
snake_case_ :Any = tokenizer.max_model_input_sizes[args.teacher_name]
# DATA LOADER #
logger.info(f"""Loading data from {args.data_file}""" )
with open(args.data_file, """rb""" ) as fp:
snake_case_ :str = pickle.load(_lowercase )
if args.mlm:
logger.info(f"""Loading token counts from {args.token_counts} (already pre-computed)""" )
with open(args.token_counts, """rb""" ) as fp:
snake_case_ :Optional[Any] = pickle.load(_lowercase )
snake_case_ :Tuple = np.maximum(_lowercase, 1 ) ** -args.mlm_smoothing
for idx in special_tok_ids.values():
snake_case_ :Optional[int] = 0.0 # do not predict special tokens
snake_case_ :int = torch.from_numpy(_lowercase )
else:
snake_case_ :List[str] = None
snake_case_ :Optional[int] = LmSeqsDataset(params=_lowercase, data=_lowercase )
logger.info("""Data loader created.""" )
# STUDENT #
logger.info(f"""Loading student config from {args.student_config}""" )
snake_case_ :List[Any] = student_config_class.from_pretrained(args.student_config )
snake_case_ :Union[str, Any] = True
if args.student_pretrained_weights is not None:
logger.info(f"""Loading pretrained weights from {args.student_pretrained_weights}""" )
snake_case_ :List[str] = student_model_class.from_pretrained(args.student_pretrained_weights, config=_lowercase )
else:
snake_case_ :Optional[int] = student_model_class(_lowercase )
if args.n_gpu > 0:
student.to(f"""cuda:{args.local_rank}""" )
logger.info("""Student loaded.""" )
# TEACHER #
snake_case_ :Dict = teacher_model_class.from_pretrained(args.teacher_name, output_hidden_states=_lowercase )
if args.n_gpu > 0:
teacher.to(f"""cuda:{args.local_rank}""" )
logger.info(f"""Teacher loaded from {args.teacher_name}.""" )
# FREEZING #
if args.freeze_pos_embs:
freeze_pos_embeddings(_lowercase, _lowercase )
if args.freeze_token_type_embds:
freeze_token_type_embeddings(_lowercase, _lowercase )
# SANITY CHECKS #
assert student.config.vocab_size == teacher.config.vocab_size
assert student.config.hidden_size == teacher.config.hidden_size
assert student.config.max_position_embeddings == teacher.config.max_position_embeddings
if args.mlm:
assert token_probs.size(0 ) == stu_architecture_config.vocab_size
# DISTILLER #
torch.cuda.empty_cache()
snake_case_ :Optional[int] = Distiller(
params=_lowercase, dataset=_lowercase, token_probs=_lowercase, student=_lowercase, teacher=_lowercase )
distiller.train()
logger.info("""Let's go get some drinks.""" )
if __name__ == "__main__":
main()
| 66 | 1 |
"""simple docstring"""
import shutil
import tempfile
import unittest
from transformers import SPIECE_UNDERLINE, BatchEncoding, MBartTokenizer, MBartTokenizerFast, is_torch_available
from transformers.testing_utils import (
get_tests_dir,
nested_simplify,
require_sentencepiece,
require_tokenizers,
require_torch,
)
from ...test_tokenization_common import TokenizerTesterMixin
__a = get_tests_dir("fixtures/test_sentencepiece.model")
if is_torch_available():
from transformers.models.mbart.modeling_mbart import shift_tokens_right
__a = 25_00_04
__a = 25_00_20
@require_sentencepiece
@require_tokenizers
class lowerCamelCase ( _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[str] = MBartTokenizer
_A : str = MBartTokenizerFast
_A : Union[str, Any] = True
_A : List[Any] = True
def lowerCAmelCase_ ( self: Optional[Any] ) -> str:
super().setUp()
# We have a SentencePiece fixture for testing
snake_case_ :Tuple = MBartTokenizer(snake_case , keep_accents=snake_case )
tokenizer.save_pretrained(self.tmpdirname )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> List[str]:
snake_case_ :str = MBartTokenizer(snake_case , keep_accents=snake_case )
snake_case_ :int = tokenizer.tokenize("""This is a test""" )
self.assertListEqual(snake_case , ["""▁This""", """▁is""", """▁a""", """▁t""", """est"""] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(snake_case ) , [value + tokenizer.fairseq_offset for value in [285, 46, 10, 170, 382]] , )
snake_case_ :Union[str, Any] = tokenizer.tokenize("""I was born in 92000, and this is falsé.""" )
self.assertListEqual(
snake_case , [
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""",
""".""",
] , )
snake_case_ :Union[str, Any] = tokenizer.convert_tokens_to_ids(snake_case )
self.assertListEqual(
snake_case , [
value + tokenizer.fairseq_offset
for value in [8, 21, 84, 55, 24, 19, 7, 2, 602, 347, 347, 347, 3, 12, 66, 46, 72, 80, 6, 2, 4]
# ^ unk: 2 + 1 = 3 unk: 2 + 1 = 3 ^
] , )
snake_case_ :Dict = tokenizer.convert_ids_to_tokens(snake_case )
self.assertListEqual(
snake_case , [
SPIECE_UNDERLINE + """I""",
SPIECE_UNDERLINE + """was""",
SPIECE_UNDERLINE + """b""",
"""or""",
"""n""",
SPIECE_UNDERLINE + """in""",
SPIECE_UNDERLINE + """""",
"""<unk>""",
"""2""",
"""0""",
"""0""",
"""0""",
""",""",
SPIECE_UNDERLINE + """and""",
SPIECE_UNDERLINE + """this""",
SPIECE_UNDERLINE + """is""",
SPIECE_UNDERLINE + """f""",
"""al""",
"""s""",
"""<unk>""",
""".""",
] , )
def lowerCAmelCase_ ( self: Optional[int] ) -> int:
if not self.test_slow_tokenizer:
# as we don't have a slow version, we can't compare the outputs between slow and fast versions
return
snake_case_ :List[str] = (self.rust_tokenizer_class, """hf-internal-testing/tiny-random-mbart""", {})
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_ :Optional[Any] = self.tokenizer_class.from_pretrained(snake_case , **snake_case )
snake_case_ :Dict = tempfile.mkdtemp()
snake_case_ :str = tokenizer_r.save_pretrained(snake_case )
snake_case_ :int = tokenizer_p.save_pretrained(snake_case )
# Checks it save with the same files + the tokenizer.json file for the fast one
self.assertTrue(any("""tokenizer.json""" in f for f in tokenizer_r_files ) )
snake_case_ :int = tuple(f for f in tokenizer_r_files if """tokenizer.json""" not in f )
self.assertSequenceEqual(snake_case , snake_case )
# Checks everything loads correctly in the same way
snake_case_ :Dict = tokenizer_r.from_pretrained(snake_case )
snake_case_ :Optional[int] = tokenizer_p.from_pretrained(snake_case )
# Check special tokens are set accordingly on Rust and Python
for key in tokenizer_pp.special_tokens_map:
self.assertTrue(hasattr(snake_case , snake_case ) )
# self.assertEqual(getattr(tokenizer_rp, key), getattr(tokenizer_pp, key))
# self.assertEqual(getattr(tokenizer_rp, key + "_id"), getattr(tokenizer_pp, key + "_id"))
shutil.rmtree(snake_case )
# Save tokenizer rust, legacy_format=True
snake_case_ :str = tempfile.mkdtemp()
snake_case_ :Union[str, Any] = tokenizer_r.save_pretrained(snake_case , legacy_format=snake_case )
snake_case_ :Any = tokenizer_p.save_pretrained(snake_case )
# Checks it save with the same files
self.assertSequenceEqual(snake_case , snake_case )
# Checks everything loads correctly in the same way
snake_case_ :Union[str, Any] = tokenizer_r.from_pretrained(snake_case )
snake_case_ :Dict = tokenizer_p.from_pretrained(snake_case )
# Check special tokens are set accordingly on Rust and Python
for key in tokenizer_pp.special_tokens_map:
self.assertTrue(hasattr(snake_case , snake_case ) )
shutil.rmtree(snake_case )
# Save tokenizer rust, legacy_format=False
snake_case_ :Tuple = tempfile.mkdtemp()
snake_case_ :Tuple = tokenizer_r.save_pretrained(snake_case , legacy_format=snake_case )
snake_case_ :int = tokenizer_p.save_pretrained(snake_case )
# Checks it saved the tokenizer.json file
self.assertTrue(any("""tokenizer.json""" in f for f in tokenizer_r_files ) )
# Checks everything loads correctly in the same way
snake_case_ :Tuple = tokenizer_r.from_pretrained(snake_case )
snake_case_ :List[Any] = tokenizer_p.from_pretrained(snake_case )
# Check special tokens are set accordingly on Rust and Python
for key in tokenizer_pp.special_tokens_map:
self.assertTrue(hasattr(snake_case , snake_case ) )
shutil.rmtree(snake_case )
@require_torch
@require_sentencepiece
@require_tokenizers
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
_A : Optional[Any] = """facebook/mbart-large-en-ro"""
_A : str = [
""" UN Chief Says There Is No Military Solution in Syria""",
""" Secretary-General Ban Ki-moon says his response to Russia's stepped up military support for Syria is that \"there is no military solution\" to the nearly five-year conflict and more weapons will only worsen the violence and misery for millions of people.""",
]
_A : List[str] = [
"""Şeful ONU declară că nu există o soluţie militară în Siria""",
"""Secretarul General Ban Ki-moon declară că răspunsul său la intensificarea sprijinului militar al Rusiei"""
""" pentru Siria este că \"nu există o soluţie militară\" la conflictul de aproape cinci ani şi că noi arme nu vor"""
""" face decât să înrăutăţească violenţele şi mizeria pentru milioane de oameni.""",
]
_A : Tuple = [8_2_7_4, 1_2_7_8_7_3, 2_5_9_1_6, 7, 8_6_2_2, 2_0_7_1, 4_3_8, 6_7_4_8_5, 5_3, 1_8_7_8_9_5, 2_3, 5_1_7_1_2, 2, EN_CODE]
@classmethod
def lowerCAmelCase_ ( cls: Optional[int] ) -> str:
snake_case_ :MBartTokenizer = MBartTokenizer.from_pretrained(
cls.checkpoint_name , src_lang="""en_XX""" , tgt_lang="""ro_RO""" )
snake_case_ :List[Any] = 1
return cls
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[Any]:
self.assertEqual(self.tokenizer.fairseq_tokens_to_ids["""ar_AR"""] , 250_001 )
self.assertEqual(self.tokenizer.fairseq_tokens_to_ids["""en_EN"""] , 250_004 )
self.assertEqual(self.tokenizer.fairseq_tokens_to_ids["""ro_RO"""] , 250_020 )
def lowerCAmelCase_ ( self: Dict ) -> Dict:
snake_case_ :List[Any] = self.tokenizer.batch_encode_plus(self.src_text ).input_ids[0]
self.assertListEqual(self.expected_src_tokens , snake_case )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Optional[int]:
self.assertIn(snake_case , self.tokenizer.all_special_ids )
snake_case_ :int = [RO_CODE, 884, 9_019, 96, 9, 916, 86_792, 36, 18_743, 15_596, 5, 2]
snake_case_ :Any = self.tokenizer.decode(snake_case , skip_special_tokens=snake_case )
snake_case_ :Optional[int] = self.tokenizer.decode(generated_ids[1:] , skip_special_tokens=snake_case )
self.assertEqual(snake_case , snake_case )
self.assertNotIn(self.tokenizer.eos_token , snake_case )
def lowerCAmelCase_ ( self: Dict ) -> str:
snake_case_ :Optional[Any] = ["""this is gunna be a long sentence """ * 20]
assert isinstance(src_text[0] , snake_case )
snake_case_ :Optional[int] = 10
snake_case_ :List[str] = self.tokenizer(snake_case , max_length=snake_case , truncation=snake_case ).input_ids[0]
self.assertEqual(ids[-2] , 2 )
self.assertEqual(ids[-1] , snake_case )
self.assertEqual(len(snake_case ) , snake_case )
def lowerCAmelCase_ ( self: Tuple ) -> Tuple:
self.assertListEqual(self.tokenizer.convert_tokens_to_ids(["""<mask>""", """ar_AR"""] ) , [250_026, 250_001] )
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[Any]:
snake_case_ :List[Any] = tempfile.mkdtemp()
snake_case_ :int = self.tokenizer.fairseq_tokens_to_ids
self.tokenizer.save_pretrained(snake_case )
snake_case_ :Dict = MBartTokenizer.from_pretrained(snake_case )
self.assertDictEqual(new_tok.fairseq_tokens_to_ids , snake_case )
@require_torch
def lowerCAmelCase_ ( self: List[str] ) -> List[str]:
snake_case_ :str = self.tokenizer(self.src_text , text_target=self.tgt_text , padding=snake_case , return_tensors="""pt""" )
snake_case_ :str = shift_tokens_right(batch["""labels"""] , self.tokenizer.pad_token_id )
# fairseq batch: https://gist.github.com/sshleifer/cba08bc2109361a74ac3760a7e30e4f4
assert batch.input_ids[1][-2:].tolist() == [2, EN_CODE]
assert batch.decoder_input_ids[1][0].tolist() == RO_CODE
assert batch.decoder_input_ids[1][-1] == 2
assert batch.labels[1][-2:].tolist() == [2, RO_CODE]
@require_torch
def lowerCAmelCase_ ( self: List[Any] ) -> Any:
snake_case_ :Union[str, Any] = self.tokenizer(
self.src_text , text_target=self.tgt_text , padding=snake_case , truncation=snake_case , max_length=len(self.expected_src_tokens ) , return_tensors="""pt""" , )
snake_case_ :int = shift_tokens_right(batch["""labels"""] , self.tokenizer.pad_token_id )
self.assertIsInstance(snake_case , snake_case )
self.assertEqual((2, 14) , batch.input_ids.shape )
self.assertEqual((2, 14) , batch.attention_mask.shape )
snake_case_ :List[Any] = batch.input_ids.tolist()[0]
self.assertListEqual(self.expected_src_tokens , snake_case )
self.assertEqual(2 , batch.decoder_input_ids[0, -1] ) # EOS
# Test that special tokens are reset
self.assertEqual(self.tokenizer.prefix_tokens , [] )
self.assertEqual(self.tokenizer.suffix_tokens , [self.tokenizer.eos_token_id, EN_CODE] )
def lowerCAmelCase_ ( self: Dict ) -> List[str]:
snake_case_ :Dict = self.tokenizer(self.src_text , padding=snake_case , truncation=snake_case , max_length=3 , return_tensors="""pt""" )
snake_case_ :Dict = self.tokenizer(
text_target=self.tgt_text , padding=snake_case , truncation=snake_case , max_length=10 , return_tensors="""pt""" )
snake_case_ :Any = targets["""input_ids"""]
snake_case_ :str = shift_tokens_right(snake_case , self.tokenizer.pad_token_id )
self.assertEqual(batch.input_ids.shape[1] , 3 )
self.assertEqual(batch.decoder_input_ids.shape[1] , 10 )
@require_torch
def lowerCAmelCase_ ( self: Dict ) -> Tuple:
snake_case_ :Optional[Any] = self.tokenizer._build_translation_inputs(
"""A test""" , return_tensors="""pt""" , src_lang="""en_XX""" , tgt_lang="""ar_AR""" )
self.assertEqual(
nested_simplify(snake_case ) , {
# A, test, EOS, en_XX
"""input_ids""": [[62, 3_034, 2, 250_004]],
"""attention_mask""": [[1, 1, 1, 1]],
# ar_AR
"""forced_bos_token_id""": 250_001,
} , )
| 66 |
"""simple docstring"""
import json
import os
import subprocess
import unittest
from ast import literal_eval
import pytest
from parameterized import parameterized, parameterized_class
from . import is_sagemaker_available
if is_sagemaker_available():
from sagemaker import Session, TrainingJobAnalytics
from sagemaker.huggingface import HuggingFace
@pytest.mark.skipif(
literal_eval(os.getenv("""TEST_SAGEMAKER""" , """False""" ) ) is not True , reason="""Skipping test because should only be run when releasing minor transformers version""" , )
@pytest.mark.usefixtures("""sm_env""" )
@parameterized_class(
[
{
"""framework""": """pytorch""",
"""script""": """run_glue_model_parallelism.py""",
"""model_name_or_path""": """roberta-large""",
"""instance_type""": """ml.p3dn.24xlarge""",
"""results""": {"""train_runtime""": 1_6_0_0, """eval_accuracy""": 0.3, """eval_loss""": 1.2},
},
{
"""framework""": """pytorch""",
"""script""": """run_glue.py""",
"""model_name_or_path""": """roberta-large""",
"""instance_type""": """ml.p3dn.24xlarge""",
"""results""": {"""train_runtime""": 1_6_0_0, """eval_accuracy""": 0.3, """eval_loss""": 1.2},
},
] )
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: Any ) -> str:
if self.framework == "pytorch":
subprocess.run(
f"""cp ./examples/pytorch/text-classification/run_glue.py {self.env.test_path}/run_glue.py""".split() , encoding="""utf-8""" , check=snake_case , )
assert hasattr(self , """env""" )
def lowerCAmelCase_ ( self: int , snake_case: Dict ) -> List[Any]:
# configuration for running training on smdistributed Model Parallel
snake_case_ :Tuple = {
"""enabled""": True,
"""processes_per_host""": 8,
}
snake_case_ :List[Any] = {
"""enabled""": True,
"""parameters""": {
"""microbatches""": 4,
"""placement_strategy""": """spread""",
"""pipeline""": """interleaved""",
"""optimize""": """speed""",
"""partitions""": 4,
"""ddp""": True,
},
}
snake_case_ :Tuple = {"""smdistributed""": {"""modelparallel""": smp_options}, """mpi""": mpi_options}
snake_case_ :Any = """trainer""" if self.script == """run_glue.py""" else """smtrainer"""
# creates estimator
return HuggingFace(
entry_point=self.script , source_dir=self.env.test_path , role=self.env.role , image_uri=self.env.image_uri , base_job_name=f"""{self.env.base_job_name}-{instance_count}-smp-{name_extension}""" , instance_count=snake_case , instance_type=self.instance_type , debugger_hook_config=snake_case , hyperparameters={
**self.env.hyperparameters,
"""model_name_or_path""": self.model_name_or_path,
"""max_steps""": 500,
} , metric_definitions=self.env.metric_definitions , distribution=snake_case , py_version="""py36""" , )
def lowerCAmelCase_ ( self: Any , snake_case: Tuple ) -> List[str]:
TrainingJobAnalytics(snake_case ).export_csv(f"""{self.env.test_path}/{job_name}_metrics.csv""" )
@parameterized.expand([(1,)] )
def lowerCAmelCase_ ( self: Dict , snake_case: Dict ) -> List[Any]:
# create estimator
snake_case_ :List[Any] = self.create_estimator(snake_case )
# run training
estimator.fit()
# result dataframe
snake_case_ :Any = TrainingJobAnalytics(estimator.latest_training_job.name ).dataframe()
# extract kpis
snake_case_ :Tuple = list(result_metrics_df[result_metrics_df.metric_name == """eval_accuracy"""]["""value"""] )
snake_case_ :Dict = list(result_metrics_df[result_metrics_df.metric_name == """eval_loss"""]["""value"""] )
# get train time from SageMaker job, this includes starting, preprocessing, stopping
snake_case_ :int = (
Session().describe_training_job(estimator.latest_training_job.name ).get("""TrainingTimeInSeconds""" , 999_999 )
)
# assert kpis
assert train_runtime <= self.results["train_runtime"]
assert all(t >= self.results["""eval_accuracy"""] for t in eval_accuracy )
assert all(t <= self.results["""eval_loss"""] for t in eval_loss )
# dump tests result into json file to share in PR
with open(f"""{estimator.latest_training_job.name}.json""" , """w""" ) as outfile:
json.dump({"""train_time""": train_runtime, """eval_accuracy""": eval_accuracy, """eval_loss""": eval_loss} , snake_case )
| 66 | 1 |
"""simple docstring"""
from __future__ import annotations
import math
def A_ ( _lowercase, _lowercase, _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
if depth < 0:
raise ValueError("""Depth cannot be less than 0""" )
if len(_lowercase ) == 0:
raise ValueError("""Scores cannot be empty""" )
if depth == height:
return scores[node_index]
if is_max:
return max(
minimax(depth + 1, node_index * 2, _lowercase, _lowercase, _lowercase ), minimax(depth + 1, node_index * 2 + 1, _lowercase, _lowercase, _lowercase ), )
return min(
minimax(depth + 1, node_index * 2, _lowercase, _lowercase, _lowercase ), minimax(depth + 1, node_index * 2 + 1, _lowercase, _lowercase, _lowercase ), )
def A_ ( ):
'''simple docstring'''
snake_case_ :List[Any] = [90, 23, 6, 33, 21, 65, 123, 34423]
snake_case_ :str = math.log(len(_lowercase ), 2 )
print("""Optimal value : """, end="""""" )
print(minimax(0, 0, _lowercase, _lowercase, _lowercase ) )
if __name__ == "__main__":
import doctest
doctest.testmod()
main()
| 66 |
"""simple docstring"""
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 lowerCamelCase :
'''simple docstring'''
def __init__( self: Dict , snake_case: Optional[Any] , snake_case: Tuple=13 , snake_case: Any=32 , snake_case: Union[str, Any]=2 , snake_case: Tuple=3 , snake_case: Union[str, Any]=16 , snake_case: Union[str, Any]=[1, 2, 1] , snake_case: Optional[Any]=[2, 2, 4] , snake_case: str=2 , snake_case: List[str]=2.0 , snake_case: Optional[int]=True , snake_case: Union[str, Any]=0.0 , snake_case: Optional[int]=0.0 , snake_case: Optional[Any]=0.1 , snake_case: List[str]="gelu" , snake_case: Any=False , snake_case: Optional[Any]=True , snake_case: Optional[int]=0.0_2 , snake_case: Any=1E-5 , snake_case: Optional[int]=True , snake_case: int=None , snake_case: Any=True , snake_case: str=10 , snake_case: Optional[Any]=8 , snake_case: Union[str, Any]=["stage1", "stage2", "stage3"] , snake_case: Tuple=[1, 2, 3] , ) -> Dict:
snake_case_ :Dict = parent
snake_case_ :List[Any] = batch_size
snake_case_ :Dict = image_size
snake_case_ :Dict = patch_size
snake_case_ :Tuple = num_channels
snake_case_ :List[Any] = embed_dim
snake_case_ :List[str] = depths
snake_case_ :str = num_heads
snake_case_ :Tuple = window_size
snake_case_ :Tuple = mlp_ratio
snake_case_ :int = qkv_bias
snake_case_ :Tuple = hidden_dropout_prob
snake_case_ :Optional[Any] = attention_probs_dropout_prob
snake_case_ :Dict = drop_path_rate
snake_case_ :Any = hidden_act
snake_case_ :Any = use_absolute_embeddings
snake_case_ :int = patch_norm
snake_case_ :List[Any] = layer_norm_eps
snake_case_ :Tuple = initializer_range
snake_case_ :str = is_training
snake_case_ :int = scope
snake_case_ :Tuple = use_labels
snake_case_ :Tuple = type_sequence_label_size
snake_case_ :str = encoder_stride
snake_case_ :List[Any] = out_features
snake_case_ :str = out_indices
def lowerCAmelCase_ ( self: Tuple ) -> Dict:
snake_case_ :Optional[int] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case_ :str = None
if self.use_labels:
snake_case_ :Optional[int] = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ :Union[str, Any] = self.get_config()
return config, pixel_values, labels
def lowerCAmelCase_ ( self: int ) -> Optional[Any]:
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 lowerCAmelCase_ ( self: List[Any] , snake_case: str , snake_case: int , snake_case: List[str] ) -> Any:
snake_case_ :Dict = MaskFormerSwinModel(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Tuple = model(snake_case )
snake_case_ :Dict = ((config.image_size // config.patch_size) ** 2) // (4 ** (len(config.depths ) - 1))
snake_case_ :Any = 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 lowerCAmelCase_ ( self: Optional[Any] , snake_case: int , snake_case: List[str] , snake_case: Tuple ) -> Union[str, Any]:
snake_case_ :Any = MaskFormerSwinBackbone(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Optional[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_ :Optional[Any] = ["""stem"""]
snake_case_ :str = MaskFormerSwinBackbone(config=snake_case )
def lowerCAmelCase_ ( self: List[str] ) -> Optional[Any]:
snake_case_ :Optional[int] = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_ :str = config_and_inputs
snake_case_ :Tuple = {"""pixel_values""": pixel_values}
return config, inputs_dict
@require_torch
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : Union[str, Any] = (
(
MaskFormerSwinModel,
MaskFormerSwinBackbone,
)
if is_torch_available()
else ()
)
_A : str = {"""feature-extraction""": MaskFormerSwinModel} if is_torch_available() else {}
_A : List[str] = False
_A : Any = False
_A : Dict = False
_A : List[Any] = False
_A : Optional[int] = False
def lowerCAmelCase_ ( self: Dict ) -> Any:
snake_case_ :str = MaskFormerSwinModelTester(self )
snake_case_ :Optional[Any] = 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 lowerCAmelCase_ ( self: List[str] ) -> Optional[int]:
pass
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Dict:
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 lowerCAmelCase_ ( self: Any ) -> Tuple:
return
def lowerCAmelCase_ ( self: Any ) -> Any:
snake_case_ :List[Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> int:
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 lowerCAmelCase_ ( self: str ) -> List[str]:
pass
@unittest.skip("""Swin does not support feedforward chunking""" )
def lowerCAmelCase_ ( self: int ) -> Optional[int]:
pass
def lowerCAmelCase_ ( self: List[str] ) -> List[Any]:
snake_case_, snake_case_ :List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :str = model_class(snake_case )
self.assertIsInstance(model.get_input_embeddings() , (nn.Module) )
snake_case_ :Dict = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(snake_case , nn.Linear ) )
def lowerCAmelCase_ ( self: Tuple ) -> Dict:
snake_case_, snake_case_ :int = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :Optional[int] = model_class(snake_case )
snake_case_ :str = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case_ :str = [*signature.parameters.keys()]
snake_case_ :str = ["""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 lowerCAmelCase_ ( self: List[Any] ) -> List[Any]:
pass
@unittest.skip(reason="""MaskFormerSwin is only used as an internal backbone""" )
def lowerCAmelCase_ ( self: Dict ) -> List[Any]:
pass
def lowerCAmelCase_ ( self: Union[str, Any] , snake_case: Union[str, Any] , snake_case: int , snake_case: Any , snake_case: List[str] ) -> str:
snake_case_ :List[str] = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :List[Any] = model(**self._prepare_for_class(snake_case , snake_case ) )
snake_case_ :Any = outputs.hidden_states
snake_case_ :Optional[int] = 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_ :str = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :int = (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 lowerCAmelCase_ ( self: List[Any] ) -> Optional[int]:
snake_case_, snake_case_ :Any = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :List[Any] = (
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_ :Tuple = 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_ :List[Any] = True
self.check_hidden_states_output(snake_case , snake_case , snake_case , snake_case )
def lowerCAmelCase_ ( self: Optional[Any] ) -> Tuple:
snake_case_, snake_case_ :int = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :List[Any] = 3
snake_case_ :List[Any] = (
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_ :Any = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :Tuple = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
snake_case_ :List[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_ :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 lowerCAmelCase_ ( self: Union[str, Any] ) -> List[str]:
pass
@unittest.skip(reason="""This will be fixed once MaskFormerSwin is replaced by native Swin""" )
def lowerCAmelCase_ ( self: List[str] ) -> str:
pass
@unittest.skip(reason="""This will be fixed once MaskFormerSwin is replaced by native Swin""" )
def lowerCAmelCase_ ( self: str ) -> List[Any]:
pass
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Optional[Any]:
snake_case_, snake_case_ :Dict = self.model_tester.prepare_config_and_inputs_for_common()
def set_nan_tensor_to_zero(snake_case: str ):
snake_case_ :Optional[int] = 0
return t
def check_equivalence(snake_case: List[Any] , snake_case: Union[str, Any] , snake_case: int , snake_case: Tuple={} ):
with torch.no_grad():
snake_case_ :List[Any] = model(**snake_case , return_dict=snake_case , **snake_case )
snake_case_ :Any = model(**snake_case , return_dict=snake_case , **snake_case ).to_tuple()
def recursive_check(snake_case: List[Any] , snake_case: int ):
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_ :int = model_class(snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Any = 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 )
snake_case_ :Tuple = self._prepare_for_class(snake_case , snake_case , return_labels=snake_case )
snake_case_ :Dict = self._prepare_for_class(snake_case , snake_case , return_labels=snake_case )
check_equivalence(snake_case , snake_case , snake_case )
snake_case_ :Tuple = self._prepare_for_class(snake_case , snake_case )
snake_case_ :Any = self._prepare_for_class(snake_case , snake_case )
check_equivalence(snake_case , snake_case , snake_case , {"""output_hidden_states""": True} )
snake_case_ :Dict = 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 lowerCamelCase ( unittest.TestCase , _lowerCAmelCase ):
'''simple docstring'''
_A : int = (MaskFormerSwinBackbone,) if is_torch_available() else ()
_A : Tuple = MaskFormerSwinConfig
def lowerCAmelCase_ ( self: List[str] ) -> Optional[int]:
snake_case_ :Optional[Any] = MaskFormerSwinModelTester(self )
def lowerCAmelCase_ ( self: int ) -> Optional[int]:
snake_case_, snake_case_ :Any = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Tuple = inputs_dict["""pixel_values"""].shape[0]
for backbone_class in self.all_model_classes:
snake_case_ :List[str] = backbone_class(snake_case )
backbone.to(snake_case )
backbone.eval()
snake_case_ :List[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_ :Union[str, Any] = 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_ :List[Any] = 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_ :List[Any] = backbone(**snake_case , output_attentions=snake_case )
self.assertIsNotNone(outputs.attentions )
| 66 | 1 |
"""simple docstring"""
import argparse
import json
import os
import pickle
import shutil
import numpy as np
import torch
from distiller import Distiller
from lm_seqs_dataset import LmSeqsDataset
from transformers import (
BertConfig,
BertForMaskedLM,
BertTokenizer,
DistilBertConfig,
DistilBertForMaskedLM,
DistilBertTokenizer,
GPTaConfig,
GPTaLMHeadModel,
GPTaTokenizer,
RobertaConfig,
RobertaForMaskedLM,
RobertaTokenizer,
)
from utils import git_log, init_gpu_params, logger, set_seed
__a = {
"distilbert": (DistilBertConfig, DistilBertForMaskedLM, DistilBertTokenizer),
"roberta": (RobertaConfig, RobertaForMaskedLM, RobertaTokenizer),
"bert": (BertConfig, BertForMaskedLM, BertTokenizer),
"gpt2": (GPTaConfig, GPTaLMHeadModel, GPTaTokenizer),
}
def A_ ( _lowercase ):
'''simple docstring'''
assert (args.mlm and args.alpha_mlm > 0.0) or (not args.mlm and args.alpha_mlm == 0.0)
assert (args.alpha_mlm > 0.0 and args.alpha_clm == 0.0) or (args.alpha_mlm == 0.0 and args.alpha_clm > 0.0)
if args.mlm:
assert os.path.isfile(args.token_counts )
assert (args.student_type in ["roberta", "distilbert"]) and (args.teacher_type in ["roberta", "bert"])
else:
assert (args.student_type in ["gpt2"]) and (args.teacher_type in ["gpt2"])
assert args.teacher_type == args.student_type or (
args.student_type == "distilbert" and args.teacher_type == "bert"
)
assert os.path.isfile(args.student_config )
if args.student_pretrained_weights is not None:
assert os.path.isfile(args.student_pretrained_weights )
if args.freeze_token_type_embds:
assert args.student_type in ["roberta"]
assert args.alpha_ce >= 0.0
assert args.alpha_mlm >= 0.0
assert args.alpha_clm >= 0.0
assert args.alpha_mse >= 0.0
assert args.alpha_cos >= 0.0
assert args.alpha_ce + args.alpha_mlm + args.alpha_clm + args.alpha_mse + args.alpha_cos > 0.0
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
if args.student_type == "roberta":
snake_case_ :Tuple = False
elif args.student_type == "gpt2":
snake_case_ :Union[str, Any] = False
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
if args.student_type == "roberta":
snake_case_ :List[str] = False
def A_ ( ):
'''simple docstring'''
snake_case_ :Union[str, Any] = argparse.ArgumentParser(description="""Training""" )
parser.add_argument("""--force""", action="""store_true""", help="""Overwrite dump_path if it already exists.""" )
parser.add_argument(
"""--dump_path""", type=_lowercase, required=_lowercase, help="""The output directory (log, checkpoints, parameters, etc.)""" )
parser.add_argument(
"""--data_file""", type=_lowercase, required=_lowercase, help="""The binarized file (tokenized + tokens_to_ids) and grouped by sequence.""", )
parser.add_argument(
"""--student_type""", type=_lowercase, choices=["""distilbert""", """roberta""", """gpt2"""], required=_lowercase, help="""The student type (DistilBERT, RoBERTa).""", )
parser.add_argument("""--student_config""", type=_lowercase, required=_lowercase, help="""Path to the student configuration.""" )
parser.add_argument(
"""--student_pretrained_weights""", default=_lowercase, type=_lowercase, help="""Load student initialization checkpoint.""" )
parser.add_argument(
"""--teacher_type""", choices=["""bert""", """roberta""", """gpt2"""], required=_lowercase, help="""Teacher type (BERT, RoBERTa).""" )
parser.add_argument("""--teacher_name""", type=_lowercase, required=_lowercase, help="""The teacher model.""" )
parser.add_argument("""--temperature""", default=2.0, type=_lowercase, help="""Temperature for the softmax temperature.""" )
parser.add_argument(
"""--alpha_ce""", default=0.5, type=_lowercase, help="""Linear weight for the distillation loss. Must be >=0.""" )
parser.add_argument(
"""--alpha_mlm""", default=0.0, type=_lowercase, help="""Linear weight for the MLM loss. Must be >=0. Should be used in conjunction with `mlm` flag.""", )
parser.add_argument("""--alpha_clm""", default=0.5, type=_lowercase, help="""Linear weight for the CLM loss. Must be >=0.""" )
parser.add_argument("""--alpha_mse""", default=0.0, type=_lowercase, help="""Linear weight of the MSE loss. Must be >=0.""" )
parser.add_argument(
"""--alpha_cos""", default=0.0, type=_lowercase, help="""Linear weight of the cosine embedding loss. Must be >=0.""" )
parser.add_argument(
"""--mlm""", action="""store_true""", help="""The LM step: MLM or CLM. If `mlm` is True, the MLM is used over CLM.""" )
parser.add_argument(
"""--mlm_mask_prop""", default=0.15, type=_lowercase, help="""Proportion of tokens for which we need to make a prediction.""", )
parser.add_argument("""--word_mask""", default=0.8, type=_lowercase, help="""Proportion of tokens to mask out.""" )
parser.add_argument("""--word_keep""", default=0.1, type=_lowercase, help="""Proportion of tokens to keep.""" )
parser.add_argument("""--word_rand""", default=0.1, type=_lowercase, help="""Proportion of tokens to randomly replace.""" )
parser.add_argument(
"""--mlm_smoothing""", default=0.7, type=_lowercase, help="""Smoothing parameter to emphasize more rare tokens (see XLM, similar to word2vec).""", )
parser.add_argument("""--token_counts""", type=_lowercase, help="""The token counts in the data_file for MLM.""" )
parser.add_argument(
"""--restrict_ce_to_mask""", action="""store_true""", help="""If true, compute the distillation loss only the [MLM] prediction distribution.""", )
parser.add_argument(
"""--freeze_pos_embs""", action="""store_true""", help="""Freeze positional embeddings during distillation. For student_type in ['roberta', 'gpt2'] only.""", )
parser.add_argument(
"""--freeze_token_type_embds""", action="""store_true""", help="""Freeze token type embeddings during distillation if existent. For student_type in ['roberta'] only.""", )
parser.add_argument("""--n_epoch""", type=_lowercase, default=3, help="""Number of pass on the whole dataset.""" )
parser.add_argument("""--batch_size""", type=_lowercase, default=5, help="""Batch size (for each process).""" )
parser.add_argument(
"""--group_by_size""", action="""store_false""", help="""If true, group sequences that have similar length into the same batch. Default is true.""", )
parser.add_argument(
"""--gradient_accumulation_steps""", type=_lowercase, default=50, help="""Gradient accumulation for larger training batches.""", )
parser.add_argument("""--warmup_prop""", default=0.05, type=_lowercase, help="""Linear warmup proportion.""" )
parser.add_argument("""--weight_decay""", default=0.0, type=_lowercase, help="""Weight decay if we apply some.""" )
parser.add_argument("""--learning_rate""", default=5e-4, type=_lowercase, help="""The initial learning rate for Adam.""" )
parser.add_argument("""--adam_epsilon""", default=1e-6, type=_lowercase, help="""Epsilon for Adam optimizer.""" )
parser.add_argument("""--max_grad_norm""", default=5.0, type=_lowercase, help="""Max gradient norm.""" )
parser.add_argument("""--initializer_range""", default=0.02, type=_lowercase, help="""Random initialization range.""" )
parser.add_argument(
"""--fp16""", action="""store_true""", help="""Whether to use 16-bit (mixed) precision (through NVIDIA apex) instead of 32-bit""", )
parser.add_argument(
"""--fp16_opt_level""", type=_lowercase, default="""O1""", help=(
"""For fp16: Apex AMP optimization level selected in ['O0', 'O1', 'O2', and 'O3']."""
"""See details at https://nvidia.github.io/apex/amp.html"""
), )
parser.add_argument("""--n_gpu""", type=_lowercase, default=1, help="""Number of GPUs in the node.""" )
parser.add_argument("""--local_rank""", type=_lowercase, default=-1, help="""Distributed training - Local rank""" )
parser.add_argument("""--seed""", type=_lowercase, default=56, help="""Random seed""" )
parser.add_argument("""--log_interval""", type=_lowercase, default=500, help="""Tensorboard logging interval.""" )
parser.add_argument("""--checkpoint_interval""", type=_lowercase, default=4000, help="""Checkpoint interval.""" )
snake_case_ :Tuple = parser.parse_args()
sanity_checks(_lowercase )
# ARGS #
init_gpu_params(_lowercase )
set_seed(_lowercase )
if args.is_master:
if os.path.exists(args.dump_path ):
if not args.force:
raise ValueError(
f"""Serialization dir {args.dump_path} already exists, but you have not precised wheter to overwrite"""
""" itUse `--force` if you want to overwrite it""" )
else:
shutil.rmtree(args.dump_path )
if not os.path.exists(args.dump_path ):
os.makedirs(args.dump_path )
logger.info(f"""Experiment will be dumped and logged in {args.dump_path}""" )
# SAVE PARAMS #
logger.info(f"""Param: {args}""" )
with open(os.path.join(args.dump_path, """parameters.json""" ), """w""" ) as f:
json.dump(vars(_lowercase ), _lowercase, indent=4 )
git_log(args.dump_path )
snake_case_, snake_case_, snake_case_ :Any = MODEL_CLASSES[args.student_type]
snake_case_, snake_case_, snake_case_ :int = MODEL_CLASSES[args.teacher_type]
# TOKENIZER #
snake_case_ :Any = teacher_tokenizer_class.from_pretrained(args.teacher_name )
snake_case_ :Optional[Any] = {}
for tok_name, tok_symbol in tokenizer.special_tokens_map.items():
snake_case_ :Union[str, Any] = tokenizer.all_special_tokens.index(_lowercase )
snake_case_ :Union[str, Any] = tokenizer.all_special_ids[idx]
logger.info(f"""Special tokens {special_tok_ids}""" )
snake_case_ :str = special_tok_ids
snake_case_ :Any = tokenizer.max_model_input_sizes[args.teacher_name]
# DATA LOADER #
logger.info(f"""Loading data from {args.data_file}""" )
with open(args.data_file, """rb""" ) as fp:
snake_case_ :str = pickle.load(_lowercase )
if args.mlm:
logger.info(f"""Loading token counts from {args.token_counts} (already pre-computed)""" )
with open(args.token_counts, """rb""" ) as fp:
snake_case_ :Optional[Any] = pickle.load(_lowercase )
snake_case_ :Tuple = np.maximum(_lowercase, 1 ) ** -args.mlm_smoothing
for idx in special_tok_ids.values():
snake_case_ :Optional[int] = 0.0 # do not predict special tokens
snake_case_ :int = torch.from_numpy(_lowercase )
else:
snake_case_ :List[str] = None
snake_case_ :Optional[int] = LmSeqsDataset(params=_lowercase, data=_lowercase )
logger.info("""Data loader created.""" )
# STUDENT #
logger.info(f"""Loading student config from {args.student_config}""" )
snake_case_ :List[Any] = student_config_class.from_pretrained(args.student_config )
snake_case_ :Union[str, Any] = True
if args.student_pretrained_weights is not None:
logger.info(f"""Loading pretrained weights from {args.student_pretrained_weights}""" )
snake_case_ :List[str] = student_model_class.from_pretrained(args.student_pretrained_weights, config=_lowercase )
else:
snake_case_ :Optional[int] = student_model_class(_lowercase )
if args.n_gpu > 0:
student.to(f"""cuda:{args.local_rank}""" )
logger.info("""Student loaded.""" )
# TEACHER #
snake_case_ :Dict = teacher_model_class.from_pretrained(args.teacher_name, output_hidden_states=_lowercase )
if args.n_gpu > 0:
teacher.to(f"""cuda:{args.local_rank}""" )
logger.info(f"""Teacher loaded from {args.teacher_name}.""" )
# FREEZING #
if args.freeze_pos_embs:
freeze_pos_embeddings(_lowercase, _lowercase )
if args.freeze_token_type_embds:
freeze_token_type_embeddings(_lowercase, _lowercase )
# SANITY CHECKS #
assert student.config.vocab_size == teacher.config.vocab_size
assert student.config.hidden_size == teacher.config.hidden_size
assert student.config.max_position_embeddings == teacher.config.max_position_embeddings
if args.mlm:
assert token_probs.size(0 ) == stu_architecture_config.vocab_size
# DISTILLER #
torch.cuda.empty_cache()
snake_case_ :Optional[int] = Distiller(
params=_lowercase, dataset=_lowercase, token_probs=_lowercase, student=_lowercase, teacher=_lowercase )
distiller.train()
logger.info("""Let's go get some drinks.""" )
if __name__ == "__main__":
main()
| 66 |
"""simple docstring"""
import gc
import math
import unittest
import torch
from diffusers import UNetaDModel
from diffusers.utils import floats_tensor, logging, slow, torch_all_close, torch_device
from diffusers.utils.testing_utils import enable_full_determinism
from .test_modeling_common import ModelTesterMixin, UNetTesterMixin
__a = logging.get_logger(__name__)
enable_full_determinism()
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[Any] = UNetaDModel
_A : Union[str, Any] = """sample"""
@property
def lowerCAmelCase_ ( self: str ) -> Tuple:
snake_case_ :List[str] = 4
snake_case_ :Tuple = 3
snake_case_ :Optional[Any] = (32, 32)
snake_case_ :str = floats_tensor((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :Union[str, Any] = torch.tensor([10] ).to(snake_case )
return {"sample": noise, "timestep": time_step}
@property
def lowerCAmelCase_ ( self: List[str] ) -> Dict:
return (3, 32, 32)
@property
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
return (3, 32, 32)
def lowerCAmelCase_ ( self: Optional[int] ) -> Dict:
snake_case_ :Any = {
"""block_out_channels""": (32, 64),
"""down_block_types""": ("""DownBlock2D""", """AttnDownBlock2D"""),
"""up_block_types""": ("""AttnUpBlock2D""", """UpBlock2D"""),
"""attention_head_dim""": 3,
"""out_channels""": 3,
"""in_channels""": 3,
"""layers_per_block""": 2,
"""sample_size""": 32,
}
snake_case_ :Tuple = self.dummy_input
return init_dict, inputs_dict
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[str] = UNetaDModel
_A : Union[str, Any] = """sample"""
@property
def lowerCAmelCase_ ( self: str ) -> str:
snake_case_ :List[str] = 4
snake_case_ :Optional[int] = 4
snake_case_ :int = (32, 32)
snake_case_ :Any = floats_tensor((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :List[Any] = torch.tensor([10] ).to(snake_case )
return {"sample": noise, "timestep": time_step}
@property
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
return (4, 32, 32)
@property
def lowerCAmelCase_ ( self: List[Any] ) -> int:
return (4, 32, 32)
def lowerCAmelCase_ ( self: Union[str, Any] ) -> List[Any]:
snake_case_ :Dict = {
"""sample_size""": 32,
"""in_channels""": 4,
"""out_channels""": 4,
"""layers_per_block""": 2,
"""block_out_channels""": (32, 64),
"""attention_head_dim""": 32,
"""down_block_types""": ("""DownBlock2D""", """DownBlock2D"""),
"""up_block_types""": ("""UpBlock2D""", """UpBlock2D"""),
}
snake_case_ :List[str] = self.dummy_input
return init_dict, inputs_dict
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[Any]:
snake_case_, snake_case_ :List[str] = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case )
self.assertIsNotNone(snake_case )
self.assertEqual(len(loading_info["""missing_keys"""] ) , 0 )
model.to(snake_case )
snake_case_ :List[str] = model(**self.dummy_input ).sample
assert image is not None, "Make sure output is not None"
@unittest.skipIf(torch_device != """cuda""" , """This test is supposed to run on GPU""" )
def lowerCAmelCase_ ( self: Tuple ) -> Dict:
snake_case_, snake_case_ :Union[str, Any] = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case )
model.to(snake_case )
snake_case_ :Union[str, Any] = model(**self.dummy_input ).sample
assert image is not None, "Make sure output is not None"
@unittest.skipIf(torch_device != """cuda""" , """This test is supposed to run on GPU""" )
def lowerCAmelCase_ ( self: str ) -> Any:
# by defautl model loading will use accelerate as `low_cpu_mem_usage=True`
snake_case_, snake_case_ :List[str] = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case )
model_accelerate.to(snake_case )
model_accelerate.eval()
snake_case_ :List[Any] = torch.randn(
1 , model_accelerate.config.in_channels , model_accelerate.config.sample_size , model_accelerate.config.sample_size , generator=torch.manual_seed(0 ) , )
snake_case_ :int = noise.to(snake_case )
snake_case_ :str = torch.tensor([10] * noise.shape[0] ).to(snake_case )
snake_case_ :Optional[int] = model_accelerate(snake_case , snake_case )["""sample"""]
# two models don't need to stay in the device at the same time
del model_accelerate
torch.cuda.empty_cache()
gc.collect()
snake_case_, snake_case_ :str = UNetaDModel.from_pretrained(
"""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case , low_cpu_mem_usage=snake_case )
model_normal_load.to(snake_case )
model_normal_load.eval()
snake_case_ :int = model_normal_load(snake_case , snake_case )["""sample"""]
assert torch_all_close(snake_case , snake_case , rtol=1E-3 )
def lowerCAmelCase_ ( self: Tuple ) -> Any:
snake_case_ :Tuple = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" )
model.eval()
model.to(snake_case )
snake_case_ :Optional[int] = torch.randn(
1 , model.config.in_channels , model.config.sample_size , model.config.sample_size , generator=torch.manual_seed(0 ) , )
snake_case_ :int = noise.to(snake_case )
snake_case_ :List[Any] = torch.tensor([10] * noise.shape[0] ).to(snake_case )
with torch.no_grad():
snake_case_ :Union[str, Any] = model(snake_case , snake_case ).sample
snake_case_ :Optional[int] = output[0, -1, -3:, -3:].flatten().cpu()
# fmt: off
snake_case_ :Dict = torch.tensor([-1_3.3_2_5_8, -2_0.1_1_0_0, -1_5.9_8_7_3, -1_7.6_6_1_7, -2_3.0_5_9_6, -1_7.9_4_1_9, -1_3.3_6_7_5, -1_6.1_8_8_9, -1_2.3_8_0_0] )
# fmt: on
self.assertTrue(torch_all_close(snake_case , snake_case , rtol=1E-3 ) )
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[Any] = UNetaDModel
_A : List[Any] = """sample"""
@property
def lowerCAmelCase_ ( self: Union[str, Any] , snake_case: int=(32, 32) ) -> Tuple:
snake_case_ :Union[str, Any] = 4
snake_case_ :Any = 3
snake_case_ :int = floats_tensor((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :Any = torch.tensor(batch_size * [10] ).to(dtype=torch.intaa , device=snake_case )
return {"sample": noise, "timestep": time_step}
@property
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Any:
return (3, 32, 32)
@property
def lowerCAmelCase_ ( self: int ) -> Tuple:
return (3, 32, 32)
def lowerCAmelCase_ ( self: List[str] ) -> Tuple:
snake_case_ :List[Any] = {
"""block_out_channels""": [32, 64, 64, 64],
"""in_channels""": 3,
"""layers_per_block""": 1,
"""out_channels""": 3,
"""time_embedding_type""": """fourier""",
"""norm_eps""": 1E-6,
"""mid_block_scale_factor""": math.sqrt(2.0 ),
"""norm_num_groups""": None,
"""down_block_types""": [
"""SkipDownBlock2D""",
"""AttnSkipDownBlock2D""",
"""SkipDownBlock2D""",
"""SkipDownBlock2D""",
],
"""up_block_types""": [
"""SkipUpBlock2D""",
"""SkipUpBlock2D""",
"""AttnSkipUpBlock2D""",
"""SkipUpBlock2D""",
],
}
snake_case_ :int = self.dummy_input
return init_dict, inputs_dict
@slow
def lowerCAmelCase_ ( self: Optional[Any] ) -> List[Any]:
snake_case_, snake_case_ :List[Any] = UNetaDModel.from_pretrained("""google/ncsnpp-celebahq-256""" , output_loading_info=snake_case )
self.assertIsNotNone(snake_case )
self.assertEqual(len(loading_info["""missing_keys"""] ) , 0 )
model.to(snake_case )
snake_case_ :Any = self.dummy_input
snake_case_ :int = floats_tensor((4, 3) + (256, 256) ).to(snake_case )
snake_case_ :int = noise
snake_case_ :int = model(**snake_case )
assert image is not None, "Make sure output is not None"
@slow
def lowerCAmelCase_ ( self: str ) -> Dict:
snake_case_ :Dict = UNetaDModel.from_pretrained("""google/ncsnpp-celebahq-256""" )
model.to(snake_case )
snake_case_ :List[str] = 4
snake_case_ :Optional[int] = 3
snake_case_ :List[str] = (256, 256)
snake_case_ :Tuple = torch.ones((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :str = torch.tensor(batch_size * [1E-4] ).to(snake_case )
with torch.no_grad():
snake_case_ :Dict = model(snake_case , snake_case ).sample
snake_case_ :int = output[0, -3:, -3:, -1].flatten().cpu()
# fmt: off
snake_case_ :Optional[Any] = torch.tensor([-4_8_4_2.8_6_9_1, -6_4_9_9.6_6_3_1, -3_8_0_0.1_9_5_3, -7_9_7_8.2_6_8_6, -1_0_9_8_0.7_1_2_9, -2_0_0_2_8.8_5_3_5, 8_1_4_8.2_8_2_2, 2_3_4_2.2_9_0_5, 5_6_7.7_6_0_8] )
# fmt: on
self.assertTrue(torch_all_close(snake_case , snake_case , rtol=1E-2 ) )
def lowerCAmelCase_ ( self: List[str] ) -> List[Any]:
snake_case_ :Optional[Any] = UNetaDModel.from_pretrained("""fusing/ncsnpp-ffhq-ve-dummy-update""" )
model.to(snake_case )
snake_case_ :Optional[int] = 4
snake_case_ :Optional[Any] = 3
snake_case_ :Optional[Any] = (32, 32)
snake_case_ :Dict = torch.ones((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :Any = torch.tensor(batch_size * [1E-4] ).to(snake_case )
with torch.no_grad():
snake_case_ :str = model(snake_case , snake_case ).sample
snake_case_ :int = output[0, -3:, -3:, -1].flatten().cpu()
# fmt: off
snake_case_ :int = torch.tensor([-0.0_3_2_5, -0.0_9_0_0, -0.0_8_6_9, -0.0_3_3_2, -0.0_7_2_5, -0.0_2_7_0, -0.0_1_0_1, 0.0_2_2_7, 0.0_2_5_6] )
# fmt: on
self.assertTrue(torch_all_close(snake_case , snake_case , rtol=1E-2 ) )
def lowerCAmelCase_ ( self: Dict ) -> Optional[Any]:
# not required for this model
pass
| 66 | 1 |
"""simple docstring"""
import argparse
import os
import torch
from transformers import FlavaImageCodebook, FlavaImageCodebookConfig
def A_ ( _lowercase, _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :Dict = s.rsplit(_lowercase, _lowercase )
return new.join(_lowercase )
def A_ ( _lowercase ):
'''simple docstring'''
return sum(param.float().sum() if """encoder.embeddings""" not in key else 0 for key, param in state_dict.items() )
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Tuple = {}
snake_case_ :Union[str, Any] = ["""group_1""", """group_2""", """group_3""", """group_4"""]
for key, value in state_dict.items():
for group_key in group_keys:
if group_key in key:
snake_case_ :int = key.replace(f"""{group_key}.""", f"""{group_key}.group.""" )
if "res_path" in key:
snake_case_ :Tuple = key.replace("""res_path.""", """res_path.path.""" )
if key.endswith(""".w""" ):
snake_case_ :List[str] = rreplace(_lowercase, """.w""", """.weight""", 1 )
if key.endswith(""".b""" ):
snake_case_ :Dict = rreplace(_lowercase, """.b""", """.bias""", 1 )
snake_case_ :Union[str, Any] = value.float()
return upgrade
@torch.no_grad()
def A_ ( _lowercase, _lowercase, _lowercase=None, _lowercase=True ):
'''simple docstring'''
from dall_e import Encoder
snake_case_ :Tuple = Encoder()
if os.path.exists(_lowercase ):
snake_case_ :Any = torch.load(_lowercase )
else:
snake_case_ :str = torch.hub.load_state_dict_from_url(_lowercase )
if isinstance(_lowercase, _lowercase ):
snake_case_ :str = ckpt.state_dict()
encoder.load_state_dict(_lowercase )
if config_path is not None:
snake_case_ :Tuple = FlavaImageCodebookConfig.from_pretrained(_lowercase )
else:
snake_case_ :int = FlavaImageCodebookConfig()
snake_case_ :List[Any] = FlavaImageCodebook(_lowercase ).eval()
snake_case_ :Union[str, Any] = encoder.state_dict()
snake_case_ :List[Any] = upgrade_state_dict(_lowercase )
hf_model.load_state_dict(_lowercase )
snake_case_ :str = hf_model.state_dict()
snake_case_ :Tuple = count_parameters(_lowercase )
snake_case_ :Optional[Any] = count_parameters(_lowercase )
assert torch.allclose(_lowercase, _lowercase, atol=1e-3 )
if save_checkpoint:
hf_model.save_pretrained(_lowercase )
else:
return hf_state_dict
if __name__ == "__main__":
__a = 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 flava checkpoint")
parser.add_argument("--config_path", default=None, type=str, help="Path to hf config.json of model to convert")
__a = parser.parse_args()
convert_dalle_checkpoint(args.checkpoint_path, args.pytorch_dump_folder_path, args.config_path)
| 66 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available, is_vision_available
__a = {
"configuration_mask2former": [
"MASK2FORMER_PRETRAINED_CONFIG_ARCHIVE_MAP",
"Mask2FormerConfig",
],
}
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["Mask2FormerImageProcessor"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"MASK2FORMER_PRETRAINED_MODEL_ARCHIVE_LIST",
"Mask2FormerForUniversalSegmentation",
"Mask2FormerModel",
"Mask2FormerPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_maskaformer import MASK2FORMER_PRETRAINED_CONFIG_ARCHIVE_MAP, MaskaFormerConfig
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .image_processing_maskaformer import MaskaFormerImageProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_maskaformer import (
MASK2FORMER_PRETRAINED_MODEL_ARCHIVE_LIST,
MaskaFormerForUniversalSegmentation,
MaskaFormerModel,
MaskaFormerPreTrainedModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure)
| 66 | 1 |
"""simple docstring"""
from collections import OrderedDict
from typing import Mapping
from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging
__a = logging.get_logger(__name__)
__a = {
"facebook/data2vec-text-base": "https://huggingface.co/data2vec/resolve/main/config.json",
}
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
_A : Dict = """data2vec-text"""
def __init__( self: int , snake_case: Tuple=30_522 , snake_case: Dict=768 , snake_case: Any=12 , snake_case: Tuple=12 , snake_case: List[str]=3_072 , snake_case: Dict="gelu" , snake_case: str=0.1 , snake_case: str=0.1 , snake_case: Optional[Any]=512 , snake_case: Any=2 , snake_case: Tuple=0.0_2 , snake_case: str=1E-12 , snake_case: Any=1 , snake_case: str=0 , snake_case: Union[str, Any]=2 , snake_case: List[Any]="absolute" , snake_case: List[str]=True , snake_case: Any=None , **snake_case: Union[str, Any] , ) -> List[Any]:
super().__init__(pad_token_id=snake_case , bos_token_id=snake_case , eos_token_id=snake_case , **snake_case )
snake_case_ :Optional[int] = vocab_size
snake_case_ :Any = hidden_size
snake_case_ :List[Any] = num_hidden_layers
snake_case_ :int = num_attention_heads
snake_case_ :List[Any] = hidden_act
snake_case_ :Any = intermediate_size
snake_case_ :int = hidden_dropout_prob
snake_case_ :Optional[Any] = attention_probs_dropout_prob
snake_case_ :Any = max_position_embeddings
snake_case_ :Dict = type_vocab_size
snake_case_ :str = initializer_range
snake_case_ :List[Any] = layer_norm_eps
snake_case_ :List[str] = position_embedding_type
snake_case_ :Any = use_cache
snake_case_ :Union[str, Any] = classifier_dropout
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
@property
def lowerCAmelCase_ ( self: Tuple ) -> Mapping[str, Mapping[int, str]]:
if self.task == "multiple-choice":
snake_case_ :int = {0: """batch""", 1: """choice""", 2: """sequence"""}
else:
snake_case_ :Tuple = {0: """batch""", 1: """sequence"""}
return OrderedDict(
[
("""input_ids""", dynamic_axis),
("""attention_mask""", dynamic_axis),
] )
| 66 |
"""simple docstring"""
import gc
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import (
AutoencoderKL,
DDIMScheduler,
StableDiffusionSAGPipeline,
UNetaDConditionModel,
)
from diffusers.utils import slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu
from ..pipeline_params import TEXT_TO_IMAGE_BATCH_PARAMS, TEXT_TO_IMAGE_IMAGE_PARAMS, TEXT_TO_IMAGE_PARAMS
from ..test_pipelines_common import PipelineLatentTesterMixin, PipelineTesterMixin
enable_full_determinism()
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : str = StableDiffusionSAGPipeline
_A : Optional[Any] = TEXT_TO_IMAGE_PARAMS
_A : Any = TEXT_TO_IMAGE_BATCH_PARAMS
_A : Tuple = TEXT_TO_IMAGE_IMAGE_PARAMS
_A : Tuple = TEXT_TO_IMAGE_IMAGE_PARAMS
_A : List[str] = False
def lowerCAmelCase_ ( self: Optional[Any] ) -> str:
torch.manual_seed(0 )
snake_case_ :Any = UNetaDConditionModel(
block_out_channels=(32, 64) , layers_per_block=2 , sample_size=32 , in_channels=4 , out_channels=4 , down_block_types=("""DownBlock2D""", """CrossAttnDownBlock2D""") , up_block_types=("""CrossAttnUpBlock2D""", """UpBlock2D""") , cross_attention_dim=32 , )
snake_case_ :Any = DDIMScheduler(
beta_start=0.0_0_0_8_5 , beta_end=0.0_1_2 , beta_schedule="""scaled_linear""" , clip_sample=snake_case , set_alpha_to_one=snake_case , )
torch.manual_seed(0 )
snake_case_ :Optional[int] = AutoencoderKL(
block_out_channels=[32, 64] , in_channels=3 , out_channels=3 , down_block_types=["""DownEncoderBlock2D""", """DownEncoderBlock2D"""] , up_block_types=["""UpDecoderBlock2D""", """UpDecoderBlock2D"""] , latent_channels=4 , )
torch.manual_seed(0 )
snake_case_ :Union[str, Any] = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1E-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1_000 , )
snake_case_ :Tuple = CLIPTextModel(snake_case )
snake_case_ :str = CLIPTokenizer.from_pretrained("""hf-internal-testing/tiny-random-clip""" )
snake_case_ :Dict = {
"""unet""": unet,
"""scheduler""": scheduler,
"""vae""": vae,
"""text_encoder""": text_encoder,
"""tokenizer""": tokenizer,
"""safety_checker""": None,
"""feature_extractor""": None,
}
return components
def lowerCAmelCase_ ( self: List[str] , snake_case: Tuple , snake_case: List[str]=0 ) -> str:
if str(snake_case ).startswith("""mps""" ):
snake_case_ :Tuple = torch.manual_seed(snake_case )
else:
snake_case_ :Optional[int] = torch.Generator(device=snake_case ).manual_seed(snake_case )
snake_case_ :Any = {
"""prompt""": """.""",
"""generator""": generator,
"""num_inference_steps""": 2,
"""guidance_scale""": 1.0,
"""sag_scale""": 1.0,
"""output_type""": """numpy""",
}
return inputs
def lowerCAmelCase_ ( self: Optional[int] ) -> str:
super().test_inference_batch_single_identical(expected_max_diff=3E-3 )
@slow
@require_torch_gpu
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: int ) -> str:
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def lowerCAmelCase_ ( self: int ) -> List[str]:
snake_case_ :Any = StableDiffusionSAGPipeline.from_pretrained("""CompVis/stable-diffusion-v1-4""" )
snake_case_ :int = sag_pipe.to(snake_case )
sag_pipe.set_progress_bar_config(disable=snake_case )
snake_case_ :Union[str, Any] = """."""
snake_case_ :str = torch.manual_seed(0 )
snake_case_ :str = sag_pipe(
[prompt] , generator=snake_case , guidance_scale=7.5 , sag_scale=1.0 , num_inference_steps=20 , output_type="""np""" )
snake_case_ :List[Any] = output.images
snake_case_ :Tuple = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
snake_case_ :List[Any] = np.array([0.1_5_6_8, 0.1_7_3_8, 0.1_6_9_5, 0.1_6_9_3, 0.1_5_0_7, 0.1_7_0_5, 0.1_5_4_7, 0.1_7_5_1, 0.1_9_4_9] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 5E-2
def lowerCAmelCase_ ( self: Dict ) -> str:
snake_case_ :Tuple = StableDiffusionSAGPipeline.from_pretrained("""stabilityai/stable-diffusion-2-1-base""" )
snake_case_ :Optional[int] = sag_pipe.to(snake_case )
sag_pipe.set_progress_bar_config(disable=snake_case )
snake_case_ :Tuple = """."""
snake_case_ :Union[str, Any] = torch.manual_seed(0 )
snake_case_ :Tuple = sag_pipe(
[prompt] , generator=snake_case , guidance_scale=7.5 , sag_scale=1.0 , num_inference_steps=20 , output_type="""np""" )
snake_case_ :Optional[int] = output.images
snake_case_ :Optional[Any] = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
snake_case_ :Tuple = np.array([0.3_4_5_9, 0.2_8_7_6, 0.2_5_3_7, 0.3_0_0_2, 0.2_6_7_1, 0.2_1_6_0, 0.3_0_2_6, 0.2_2_6_2, 0.2_3_7_1] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 5E-2
def lowerCAmelCase_ ( self: List[str] ) -> List[str]:
snake_case_ :Optional[int] = StableDiffusionSAGPipeline.from_pretrained("""stabilityai/stable-diffusion-2-1-base""" )
snake_case_ :int = sag_pipe.to(snake_case )
sag_pipe.set_progress_bar_config(disable=snake_case )
snake_case_ :Tuple = """."""
snake_case_ :Optional[int] = torch.manual_seed(0 )
snake_case_ :List[str] = sag_pipe(
[prompt] , width=768 , height=512 , generator=snake_case , guidance_scale=7.5 , sag_scale=1.0 , num_inference_steps=20 , output_type="""np""" , )
snake_case_ :Optional[Any] = output.images
assert image.shape == (1, 512, 768, 3)
| 66 | 1 |
"""simple docstring"""
import math
import time
from transformers import Trainer, is_torch_tpu_available
from transformers.trainer_utils import PredictionOutput, speed_metrics
if is_torch_tpu_available(check_device=False):
import torch_xla.core.xla_model as xm
import torch_xla.debug.metrics as met
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
def __init__( self: Optional[int] , *snake_case: Union[str, Any] , snake_case: Dict=None , snake_case: Optional[Any]=None , **snake_case: Tuple ) -> str:
super().__init__(*snake_case , **snake_case )
snake_case_ :List[Any] = eval_examples
snake_case_ :Optional[Any] = post_process_function
def lowerCAmelCase_ ( self: List[str] , snake_case: str=None , snake_case: List[Any]=None , snake_case: Tuple=None , snake_case: str = "eval" ) -> Union[str, Any]:
snake_case_ :Dict = self.eval_dataset if eval_dataset is None else eval_dataset
snake_case_ :str = self.get_eval_dataloader(snake_case )
snake_case_ :Union[str, Any] = self.eval_examples if eval_examples is None else eval_examples
# Temporarily disable metric computation, we will do it in the loop here.
snake_case_ :str = self.compute_metrics
snake_case_ :List[str] = None
snake_case_ :Any = self.prediction_loop if self.args.use_legacy_prediction_loop else self.evaluation_loop
snake_case_ :Optional[int] = time.time()
try:
snake_case_ :Optional[Any] = eval_loop(
snake_case , description="""Evaluation""" , prediction_loss_only=True if compute_metrics is None else None , ignore_keys=snake_case , metric_key_prefix=snake_case , )
finally:
snake_case_ :Optional[int] = compute_metrics
snake_case_ :Dict = self.args.eval_batch_size * self.args.world_size
if f"""{metric_key_prefix}_jit_compilation_time""" in output.metrics:
start_time += output.metrics[f"""{metric_key_prefix}_jit_compilation_time"""]
output.metrics.update(
speed_metrics(
snake_case , snake_case , num_samples=output.num_samples , num_steps=math.ceil(output.num_samples / total_batch_size ) , ) )
if self.post_process_function is not None and self.compute_metrics is not None and self.args.should_save:
# Only the main node write the results by default
snake_case_ :str = self.post_process_function(snake_case , snake_case , output.predictions )
snake_case_ :str = self.compute_metrics(snake_case )
# Prefix all keys with metric_key_prefix + '_'
for key in list(metrics.keys() ):
if not key.startswith(f"""{metric_key_prefix}_""" ):
snake_case_ :Tuple = metrics.pop(snake_case )
metrics.update(output.metrics )
else:
snake_case_ :int = output.metrics
if self.args.should_log:
# Only the main node log the results by default
self.log(snake_case )
if self.args.tpu_metrics_debug or self.args.debug:
# tpu-comment: Logging debug metrics for PyTorch/XLA (compile, execute times, ops, etc.)
xm.master_print(met.metrics_report() )
snake_case_ :Union[str, Any] = self.callback_handler.on_evaluate(self.args , self.state , self.control , snake_case )
return metrics
def lowerCAmelCase_ ( self: int , snake_case: Optional[Any] , snake_case: int , snake_case: Optional[Any]=None , snake_case: str = "test" ) -> Dict:
snake_case_ :List[str] = self.get_test_dataloader(snake_case )
# Temporarily disable metric computation, we will do it in the loop here.
snake_case_ :Optional[Any] = self.compute_metrics
snake_case_ :int = None
snake_case_ :List[Any] = self.prediction_loop if self.args.use_legacy_prediction_loop else self.evaluation_loop
snake_case_ :Optional[Any] = time.time()
try:
snake_case_ :Dict = eval_loop(
snake_case , description="""Prediction""" , prediction_loss_only=True if compute_metrics is None else None , ignore_keys=snake_case , metric_key_prefix=snake_case , )
finally:
snake_case_ :Any = compute_metrics
snake_case_ :int = self.args.eval_batch_size * self.args.world_size
if f"""{metric_key_prefix}_jit_compilation_time""" in output.metrics:
start_time += output.metrics[f"""{metric_key_prefix}_jit_compilation_time"""]
output.metrics.update(
speed_metrics(
snake_case , snake_case , num_samples=output.num_samples , num_steps=math.ceil(output.num_samples / total_batch_size ) , ) )
if self.post_process_function is None or self.compute_metrics is None:
return output
snake_case_ :Optional[Any] = self.post_process_function(snake_case , snake_case , output.predictions , """predict""" )
snake_case_ :Optional[Any] = self.compute_metrics(snake_case )
# Prefix all keys with metric_key_prefix + '_'
for key in list(metrics.keys() ):
if not key.startswith(f"""{metric_key_prefix}_""" ):
snake_case_ :Union[str, Any] = metrics.pop(snake_case )
metrics.update(output.metrics )
return PredictionOutput(predictions=predictions.predictions , label_ids=predictions.label_ids , metrics=snake_case )
| 66 |
"""simple docstring"""
from __future__ import annotations
from collections import Counter
from random import random
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Tuple ) -> Optional[Any]:
snake_case_ :Optional[int] = {}
def lowerCAmelCase_ ( self: Dict , snake_case: str ) -> None:
snake_case_ :str = {}
def lowerCAmelCase_ ( self: Optional[int] , snake_case: str , snake_case: str , snake_case: float ) -> None:
if nodea not in self.connections:
self.add_node(snake_case )
if nodea not in self.connections:
self.add_node(snake_case )
snake_case_ :Dict = probability
def lowerCAmelCase_ ( self: List[Any] ) -> list[str]:
return list(self.connections )
def lowerCAmelCase_ ( self: Any , snake_case: str ) -> str:
snake_case_ :Optional[Any] = 0
snake_case_ :List[str] = random()
for dest in self.connections[node]:
current_probability += self.connections[node][dest]
if current_probability > random_value:
return dest
return ""
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :List[str] = MarkovChainGraphUndirectedUnweighted()
for nodea, nodea, probability in transitions:
graph.add_transition_probability(_lowercase, _lowercase, _lowercase )
snake_case_ :int = Counter(graph.get_nodes() )
snake_case_ :Optional[Any] = start
for _ in range(_lowercase ):
snake_case_ :Tuple = graph.transition(_lowercase )
visited[node] += 1
return visited
if __name__ == "__main__":
import doctest
doctest.testmod()
| 66 | 1 |
"""simple docstring"""
import argparse
import glob
import logging
import os
import sys
import time
from collections import defaultdict
from pathlib import Path
from typing import Dict, List, Tuple
import numpy as np
import pytorch_lightning as pl
import torch
from callbacks import SeqaSeqLoggingCallback, get_checkpoint_callback, get_early_stopping_callback
from torch import nn
from torch.utils.data import DataLoader
from transformers import MBartTokenizer, TaForConditionalGeneration
from transformers.models.bart.modeling_bart import shift_tokens_right
from utils import (
ROUGE_KEYS,
LegacySeqaSeqDataset,
SeqaSeqDataset,
assert_all_frozen,
calculate_bleu,
calculate_rouge,
check_output_dir,
flatten_list,
freeze_embeds,
freeze_params,
get_git_info,
label_smoothed_nll_loss,
lmap,
pickle_save,
save_git_info,
save_json,
use_task_specific_params,
)
# need the parent dir module
sys.path.insert(2, str(Path(__file__).resolve().parents[1]))
from lightning_base import BaseTransformer, add_generic_args, generic_train # noqa
__a = logging.getLogger(__name__)
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
_A : str = """summarization"""
_A : Union[str, Any] = ["""loss"""]
_A : Dict = ROUGE_KEYS
_A : Tuple = """rouge2"""
def __init__( self: List[Any] , snake_case: List[str] , **snake_case: Tuple ) -> str:
if hparams.sortish_sampler and hparams.gpus > 1:
snake_case_ :List[str] = False
elif hparams.max_tokens_per_batch is not None:
if hparams.gpus > 1:
raise NotImplementedError("""Dynamic Batch size does not work for multi-gpu training""" )
if hparams.sortish_sampler:
raise ValueError("""--sortish_sampler and --max_tokens_per_batch may not be used simultaneously""" )
super().__init__(snake_case , num_labels=snake_case , mode=self.mode , **snake_case )
use_task_specific_params(self.model , """summarization""" )
save_git_info(self.hparams.output_dir )
snake_case_ :Dict = Path(self.output_dir ) / """metrics.json"""
snake_case_ :Dict = Path(self.output_dir ) / """hparams.pkl"""
pickle_save(self.hparams , self.hparams_save_path )
snake_case_ :Dict = 0
snake_case_ :List[Any] = defaultdict(snake_case )
snake_case_ :List[str] = self.config.model_type
snake_case_ :Any = self.config.tgt_vocab_size if self.model_type == """fsmt""" else self.config.vocab_size
snake_case_ :dict = {
"data_dir": self.hparams.data_dir,
"max_source_length": self.hparams.max_source_length,
"prefix": self.model.config.prefix or "",
}
snake_case_ :List[Any] = {
"""train""": self.hparams.n_train,
"""val""": self.hparams.n_val,
"""test""": self.hparams.n_test,
}
snake_case_ :Tuple = {k: v if v >= 0 else None for k, v in n_observations_per_split.items()}
snake_case_ :int = {
"""train""": self.hparams.max_target_length,
"""val""": self.hparams.val_max_target_length,
"""test""": self.hparams.test_max_target_length,
}
assert self.target_lens["train"] <= self.target_lens["val"], f"""target_lens: {self.target_lens}"""
assert self.target_lens["train"] <= self.target_lens["test"], f"""target_lens: {self.target_lens}"""
if self.hparams.freeze_embeds:
freeze_embeds(self.model )
if self.hparams.freeze_encoder:
freeze_params(self.model.get_encoder() )
assert_all_frozen(self.model.get_encoder() )
snake_case_ :Dict = get_git_info()["""repo_sha"""]
snake_case_ :Optional[int] = hparams.num_workers
snake_case_ :Dict = None # default to config
if self.model.config.decoder_start_token_id is None and isinstance(self.tokenizer , snake_case ):
snake_case_ :Optional[Any] = self.tokenizer.lang_code_to_id[hparams.tgt_lang]
snake_case_ :Any = self.decoder_start_token_id
snake_case_ :Optional[Any] = (
SeqaSeqDataset if hasattr(self.tokenizer , """prepare_seq2seq_batch""" ) else LegacySeqaSeqDataset
)
snake_case_ :Any = False
snake_case_ :Any = self.model.config.num_beams if self.hparams.eval_beams is None else self.hparams.eval_beams
if self.hparams.eval_max_gen_length is not None:
snake_case_ :Union[str, Any] = self.hparams.eval_max_gen_length
else:
snake_case_ :Optional[Any] = self.model.config.max_length
snake_case_ :Tuple = self.default_val_metric if self.hparams.val_metric is None else self.hparams.val_metric
def lowerCAmelCase_ ( self: str , snake_case: Dict[str, torch.Tensor] ) -> Dict[str, List[str]]:
snake_case_ :List[str] = {
k: self.tokenizer.batch_decode(v.tolist() ) if """mask""" not in k else v.shape for k, v in batch.items()
}
save_json(snake_case , Path(self.output_dir ) / """text_batch.json""" )
save_json({k: v.tolist() for k, v in batch.items()} , Path(self.output_dir ) / """tok_batch.json""" )
snake_case_ :int = True
return readable_batch
def lowerCAmelCase_ ( self: str , snake_case: Dict , **snake_case: List[Any] ) -> List[Any]:
return self.model(snake_case , **snake_case )
def lowerCAmelCase_ ( self: List[str] , snake_case: List[int] ) -> Dict:
snake_case_ :Tuple = self.tokenizer.batch_decode(
snake_case , skip_special_tokens=snake_case , clean_up_tokenization_spaces=snake_case )
return lmap(str.strip , snake_case )
def lowerCAmelCase_ ( self: str , snake_case: dict ) -> Tuple:
snake_case_ :Tuple = self.tokenizer.pad_token_id
snake_case_, snake_case_ :List[str] = batch["""input_ids"""], batch["""attention_mask"""]
snake_case_ :List[Any] = batch["""labels"""]
if isinstance(self.model , snake_case ):
snake_case_ :Dict = self.model._shift_right(snake_case )
else:
snake_case_ :List[Any] = shift_tokens_right(snake_case , snake_case )
if not self.already_saved_batch: # This would be slightly better if it only happened on rank zero
snake_case_ :Optional[Any] = decoder_input_ids
self.save_readable_batch(snake_case )
snake_case_ :List[Any] = self(snake_case , attention_mask=snake_case , decoder_input_ids=snake_case , use_cache=snake_case )
snake_case_ :Dict = outputs["""logits"""]
if self.hparams.label_smoothing == 0:
# Same behavior as modeling_bart.py, besides ignoring pad_token_id
snake_case_ :Optional[int] = nn.CrossEntropyLoss(ignore_index=snake_case )
assert lm_logits.shape[-1] == self.vocab_size
snake_case_ :Optional[Any] = ce_loss_fct(lm_logits.view(-1 , lm_logits.shape[-1] ) , tgt_ids.view(-1 ) )
else:
snake_case_ :Dict = nn.functional.log_softmax(snake_case , dim=-1 )
snake_case_, snake_case_ :Dict = label_smoothed_nll_loss(
snake_case , snake_case , self.hparams.label_smoothing , ignore_index=snake_case )
return (loss,)
@property
def lowerCAmelCase_ ( self: List[Any] ) -> int:
return self.tokenizer.pad_token_id
def lowerCAmelCase_ ( self: Any , snake_case: Union[str, Any] , snake_case: int ) -> Dict:
snake_case_ :Optional[Any] = self._step(snake_case )
snake_case_ :Optional[Any] = dict(zip(self.loss_names , snake_case ) )
# tokens per batch
snake_case_ :List[Any] = batch["""input_ids"""].ne(self.pad ).sum() + batch["""labels"""].ne(self.pad ).sum()
snake_case_ :List[Any] = batch["""input_ids"""].shape[0]
snake_case_ :str = batch["""input_ids"""].eq(self.pad ).sum()
snake_case_ :Optional[Any] = batch["""input_ids"""].eq(self.pad ).float().mean()
# TODO(SS): make a wandb summary metric for this
return {"loss": loss_tensors[0], "log": logs}
def lowerCAmelCase_ ( self: Union[str, Any] , snake_case: str , snake_case: int ) -> Dict:
return self._generative_step(snake_case )
def lowerCAmelCase_ ( self: Tuple , snake_case: str , snake_case: List[Any]="val" ) -> Dict:
self.step_count += 1
snake_case_ :int = {k: torch.stack([x[k] for x in outputs] ).mean() for k in self.loss_names}
snake_case_ :List[Any] = losses["""loss"""]
snake_case_ :Tuple = {
k: np.array([x[k] for x in outputs] ).mean() for k in self.metric_names + ["""gen_time""", """gen_len"""]
}
snake_case_ :Optional[Any] = (
generative_metrics[self.val_metric] if self.val_metric in generative_metrics else losses[self.val_metric]
)
snake_case_ :torch.FloatTensor = torch.tensor(snake_case ).type_as(snake_case )
generative_metrics.update({k: v.item() for k, v in losses.items()} )
losses.update(snake_case )
snake_case_ :List[str] = {f"""{prefix}_avg_{k}""": x for k, x in losses.items()}
snake_case_ :Optional[int] = self.step_count
self.metrics[prefix].append(snake_case ) # callback writes this to self.metrics_save_path
snake_case_ :Any = flatten_list([x["""preds"""] for x in outputs] )
return {
"log": all_metrics,
"preds": preds,
f"""{prefix}_loss""": loss,
f"""{prefix}_{self.val_metric}""": metric_tensor,
}
def lowerCAmelCase_ ( self: int , snake_case: str , snake_case: Optional[int] ) -> Dict:
return calculate_rouge(snake_case , snake_case )
def lowerCAmelCase_ ( self: Optional[int] , snake_case: dict ) -> dict:
snake_case_ :str = time.time()
# parser.add_argument('--eval_max_gen_length', type=int, default=None, help='never generate more than n tokens')
snake_case_ :Tuple = self.model.generate(
batch["""input_ids"""] , attention_mask=batch["""attention_mask"""] , use_cache=snake_case , decoder_start_token_id=self.decoder_start_token_id , num_beams=self.eval_beams , max_length=self.eval_max_length , )
snake_case_ :List[str] = (time.time() - ta) / batch["""input_ids"""].shape[0]
snake_case_ :List[str] = self.ids_to_clean_text(snake_case )
snake_case_ :List[str] = self.ids_to_clean_text(batch["""labels"""] )
snake_case_ :Dict = self._step(snake_case )
snake_case_ :Optional[int] = dict(zip(self.loss_names , snake_case ) )
snake_case_ :Dict = self.calc_generative_metrics(snake_case , snake_case )
snake_case_ :Any = np.mean(lmap(snake_case , snake_case ) )
base_metrics.update(gen_time=snake_case , gen_len=snake_case , preds=snake_case , target=snake_case , **snake_case )
return base_metrics
def lowerCAmelCase_ ( self: Optional[Any] , snake_case: str , snake_case: List[Any] ) -> Optional[Any]:
return self._generative_step(snake_case )
def lowerCAmelCase_ ( self: str , snake_case: Optional[int] ) -> List[Any]:
return self.validation_epoch_end(snake_case , prefix="""test""" )
def lowerCAmelCase_ ( self: Any , snake_case: Optional[int] ) -> SeqaSeqDataset:
snake_case_ :Optional[Any] = self.n_obs[type_path]
snake_case_ :Optional[int] = self.target_lens[type_path]
snake_case_ :List[str] = self.dataset_class(
self.tokenizer , type_path=snake_case , n_obs=snake_case , max_target_length=snake_case , **self.dataset_kwargs , )
return dataset
def lowerCAmelCase_ ( self: List[Any] , snake_case: str , snake_case: int , snake_case: bool = False ) -> DataLoader:
snake_case_ :Any = self.get_dataset(snake_case )
if self.hparams.sortish_sampler and type_path != "test" and type_path != "val":
snake_case_ :Optional[int] = dataset.make_sortish_sampler(snake_case , distributed=self.hparams.gpus > 1 )
return DataLoader(
snake_case , batch_size=snake_case , collate_fn=dataset.collate_fn , shuffle=snake_case , num_workers=self.num_workers , sampler=snake_case , )
elif self.hparams.max_tokens_per_batch is not None and type_path != "test" and type_path != "val":
snake_case_ :Union[str, Any] = dataset.make_dynamic_sampler(
self.hparams.max_tokens_per_batch , distributed=self.hparams.gpus > 1 )
return DataLoader(
snake_case , batch_sampler=snake_case , collate_fn=dataset.collate_fn , num_workers=self.num_workers , )
else:
return DataLoader(
snake_case , batch_size=snake_case , collate_fn=dataset.collate_fn , shuffle=snake_case , num_workers=self.num_workers , sampler=snake_case , )
def lowerCAmelCase_ ( self: List[Any] ) -> DataLoader:
snake_case_ :Any = self.get_dataloader("""train""" , batch_size=self.hparams.train_batch_size , shuffle=snake_case )
return dataloader
def lowerCAmelCase_ ( self: Union[str, Any] ) -> DataLoader:
return self.get_dataloader("""val""" , batch_size=self.hparams.eval_batch_size )
def lowerCAmelCase_ ( self: Dict ) -> DataLoader:
return self.get_dataloader("""test""" , batch_size=self.hparams.eval_batch_size )
@staticmethod
def lowerCAmelCase_ ( snake_case: List[Any] , snake_case: List[str] ) -> List[Any]:
BaseTransformer.add_model_specific_args(snake_case , snake_case )
add_generic_args(snake_case , snake_case )
parser.add_argument(
"""--max_source_length""" , default=1_024 , type=snake_case , help=(
"""The maximum total input sequence length after tokenization. Sequences longer """
"""than this will be truncated, sequences shorter will be padded."""
) , )
parser.add_argument(
"""--max_target_length""" , default=56 , type=snake_case , help=(
"""The maximum total input sequence length after tokenization. Sequences longer """
"""than this will be truncated, sequences shorter will be padded."""
) , )
parser.add_argument(
"""--val_max_target_length""" , default=142 , type=snake_case , help=(
"""The maximum total input sequence length after tokenization. Sequences longer """
"""than this will be truncated, sequences shorter will be padded."""
) , )
parser.add_argument(
"""--test_max_target_length""" , default=142 , type=snake_case , help=(
"""The maximum total input sequence length after tokenization. Sequences longer """
"""than this will be truncated, sequences shorter will be padded."""
) , )
parser.add_argument("""--freeze_encoder""" , action="""store_true""" )
parser.add_argument("""--freeze_embeds""" , action="""store_true""" )
parser.add_argument("""--sortish_sampler""" , action="""store_true""" , default=snake_case )
parser.add_argument("""--overwrite_output_dir""" , action="""store_true""" , default=snake_case )
parser.add_argument("""--max_tokens_per_batch""" , type=snake_case , default=snake_case )
parser.add_argument("""--logger_name""" , type=snake_case , choices=["""default""", """wandb""", """wandb_shared"""] , default="""default""" )
parser.add_argument("""--n_train""" , type=snake_case , default=-1 , required=snake_case , help="""# examples. -1 means use all.""" )
parser.add_argument("""--n_val""" , type=snake_case , default=500 , required=snake_case , help="""# examples. -1 means use all.""" )
parser.add_argument("""--n_test""" , type=snake_case , default=-1 , required=snake_case , help="""# examples. -1 means use all.""" )
parser.add_argument(
"""--task""" , type=snake_case , default="""summarization""" , required=snake_case , help="""# examples. -1 means use all.""" )
parser.add_argument("""--label_smoothing""" , type=snake_case , default=0.0 , required=snake_case )
parser.add_argument("""--src_lang""" , type=snake_case , default="""""" , required=snake_case )
parser.add_argument("""--tgt_lang""" , type=snake_case , default="""""" , required=snake_case )
parser.add_argument("""--eval_beams""" , type=snake_case , default=snake_case , required=snake_case )
parser.add_argument(
"""--val_metric""" , type=snake_case , default=snake_case , required=snake_case , choices=["""bleu""", """rouge2""", """loss""", None] )
parser.add_argument("""--eval_max_gen_length""" , type=snake_case , default=snake_case , help="""never generate more than n tokens""" )
parser.add_argument("""--save_top_k""" , type=snake_case , default=1 , required=snake_case , help="""How many checkpoints to save""" )
parser.add_argument(
"""--early_stopping_patience""" , type=snake_case , default=-1 , required=snake_case , help=(
"""-1 means never early stop. early_stopping_patience is measured in validation checks, not epochs. So"""
""" val_check_interval will effect it."""
) , )
return parser
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
_A : Any = """translation"""
_A : Tuple = ["""loss"""]
_A : List[Any] = ["""bleu"""]
_A : Tuple = """bleu"""
def __init__( self: str , snake_case: Tuple , **snake_case: int ) -> List[Any]:
super().__init__(snake_case , **snake_case )
snake_case_ :Dict = hparams.src_lang
snake_case_ :List[Any] = hparams.tgt_lang
def lowerCAmelCase_ ( self: Any , snake_case: Dict , snake_case: Optional[int] ) -> dict:
return calculate_bleu(snake_case , snake_case )
def A_ ( _lowercase, _lowercase=None ):
'''simple docstring'''
Path(args.output_dir ).mkdir(exist_ok=_lowercase )
check_output_dir(_lowercase, expected_items=3 )
if model is None:
if "summarization" in args.task:
snake_case_ :SummarizationModule = SummarizationModule(_lowercase )
else:
snake_case_ :SummarizationModule = TranslationModule(_lowercase )
snake_case_ :Tuple = Path(args.data_dir ).name
if (
args.logger_name == "default"
or args.fast_dev_run
or str(args.output_dir ).startswith("""/tmp""" )
or str(args.output_dir ).startswith("""/var""" )
):
snake_case_ :List[str] = True # don't pollute wandb logs unnecessarily
elif args.logger_name == "wandb":
from pytorch_lightning.loggers import WandbLogger
snake_case_ :List[Any] = os.environ.get("""WANDB_PROJECT""", _lowercase )
snake_case_ :Optional[int] = WandbLogger(name=model.output_dir.name, project=_lowercase )
elif args.logger_name == "wandb_shared":
from pytorch_lightning.loggers import WandbLogger
snake_case_ :Union[str, Any] = WandbLogger(name=model.output_dir.name, project=f"""hf_{dataset}""" )
if args.early_stopping_patience >= 0:
snake_case_ :Any = get_early_stopping_callback(model.val_metric, args.early_stopping_patience )
else:
snake_case_ :List[Any] = False
snake_case_ :int = args.val_metric == """loss"""
snake_case_ :pl.Trainer = generic_train(
_lowercase, _lowercase, logging_callback=SeqaSeqLoggingCallback(), checkpoint_callback=get_checkpoint_callback(
args.output_dir, model.val_metric, args.save_top_k, _lowercase ), early_stopping_callback=_lowercase, logger=_lowercase, )
pickle_save(model.hparams, model.output_dir / """hparams.pkl""" )
if not args.do_predict:
return model
snake_case_ :Tuple = """"""
snake_case_ :Union[str, Any] = sorted(glob.glob(os.path.join(args.output_dir, """*.ckpt""" ), recursive=_lowercase ) )
if checkpoints:
snake_case_ :Any = checkpoints[-1]
snake_case_ :int = checkpoints[-1]
trainer.logger.log_hyperparams(model.hparams )
# test() without a model tests using the best checkpoint automatically
trainer.test()
return model
if __name__ == "__main__":
__a = argparse.ArgumentParser()
__a = pl.Trainer.add_argparse_args(parser)
__a = SummarizationModule.add_model_specific_args(parser, os.getcwd())
__a = parser.parse_args()
main(args)
| 66 |
"""simple docstring"""
import argparse
import collections
import os
import re
import tempfile
import pandas as pd
from datasets import Dataset
from huggingface_hub import hf_hub_download, upload_folder
from transformers.utils import direct_transformers_import
# All paths are set with the intent you should run this script from the root of the repo with the command
# python utils/update_metadata.py
__a = "src/transformers"
# This is to make sure the transformers module imported is the one in the repo.
__a = direct_transformers_import(TRANSFORMERS_PATH)
# Regexes that match TF/Flax/PT model names.
__a = re.compile(r"TF(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
__a = re.compile(r"Flax(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
# Will match any TF or Flax model too so need to be in an else branch afterthe two previous regexes.
__a = re.compile(r"(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
# Fill this with tuples (pipeline_tag, model_mapping, auto_model)
__a = [
("pretraining", "MODEL_FOR_PRETRAINING_MAPPING_NAMES", "AutoModelForPreTraining"),
("feature-extraction", "MODEL_MAPPING_NAMES", "AutoModel"),
("audio-classification", "MODEL_FOR_AUDIO_CLASSIFICATION_MAPPING_NAMES", "AutoModelForAudioClassification"),
("text-generation", "MODEL_FOR_CAUSAL_LM_MAPPING_NAMES", "AutoModelForCausalLM"),
("automatic-speech-recognition", "MODEL_FOR_CTC_MAPPING_NAMES", "AutoModelForCTC"),
("image-classification", "MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES", "AutoModelForImageClassification"),
("image-segmentation", "MODEL_FOR_IMAGE_SEGMENTATION_MAPPING_NAMES", "AutoModelForImageSegmentation"),
("fill-mask", "MODEL_FOR_MASKED_LM_MAPPING_NAMES", "AutoModelForMaskedLM"),
("object-detection", "MODEL_FOR_OBJECT_DETECTION_MAPPING_NAMES", "AutoModelForObjectDetection"),
(
"zero-shot-object-detection",
"MODEL_FOR_ZERO_SHOT_OBJECT_DETECTION_MAPPING_NAMES",
"AutoModelForZeroShotObjectDetection",
),
("question-answering", "MODEL_FOR_QUESTION_ANSWERING_MAPPING_NAMES", "AutoModelForQuestionAnswering"),
("text2text-generation", "MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING_NAMES", "AutoModelForSeq2SeqLM"),
("text-classification", "MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES", "AutoModelForSequenceClassification"),
("automatic-speech-recognition", "MODEL_FOR_SPEECH_SEQ_2_SEQ_MAPPING_NAMES", "AutoModelForSpeechSeq2Seq"),
(
"table-question-answering",
"MODEL_FOR_TABLE_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForTableQuestionAnswering",
),
("token-classification", "MODEL_FOR_TOKEN_CLASSIFICATION_MAPPING_NAMES", "AutoModelForTokenClassification"),
("multiple-choice", "MODEL_FOR_MULTIPLE_CHOICE_MAPPING_NAMES", "AutoModelForMultipleChoice"),
(
"next-sentence-prediction",
"MODEL_FOR_NEXT_SENTENCE_PREDICTION_MAPPING_NAMES",
"AutoModelForNextSentencePrediction",
),
(
"audio-frame-classification",
"MODEL_FOR_AUDIO_FRAME_CLASSIFICATION_MAPPING_NAMES",
"AutoModelForAudioFrameClassification",
),
("audio-xvector", "MODEL_FOR_AUDIO_XVECTOR_MAPPING_NAMES", "AutoModelForAudioXVector"),
(
"document-question-answering",
"MODEL_FOR_DOCUMENT_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForDocumentQuestionAnswering",
),
(
"visual-question-answering",
"MODEL_FOR_VISUAL_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForVisualQuestionAnswering",
),
("image-to-text", "MODEL_FOR_FOR_VISION_2_SEQ_MAPPING_NAMES", "AutoModelForVision2Seq"),
(
"zero-shot-image-classification",
"MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING_NAMES",
"AutoModelForZeroShotImageClassification",
),
("depth-estimation", "MODEL_FOR_DEPTH_ESTIMATION_MAPPING_NAMES", "AutoModelForDepthEstimation"),
("video-classification", "MODEL_FOR_VIDEO_CLASSIFICATION_MAPPING_NAMES", "AutoModelForVideoClassification"),
("mask-generation", "MODEL_FOR_MASK_GENERATION_MAPPING_NAMES", "AutoModelForMaskGeneration"),
]
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Any = re.finditer(""".+?(?:(?<=[a-z])(?=[A-Z])|(?<=[A-Z])(?=[A-Z][a-z])|$)""", _lowercase )
return [m.group(0 ) for m in matches]
def A_ ( ):
'''simple docstring'''
snake_case_ :int = transformers_module.models.auto.configuration_auto.CONFIG_MAPPING_NAMES
snake_case_ :Dict = {
config.replace("""Config""", """""" ): model_type for model_type, config in config_maping_names.items()
}
# Dictionaries flagging if each model prefix has a backend in PT/TF/Flax.
snake_case_ :Optional[Any] = collections.defaultdict(_lowercase )
snake_case_ :int = collections.defaultdict(_lowercase )
snake_case_ :List[str] = collections.defaultdict(_lowercase )
# Let's lookup through all transformers object (once) and find if models are supported by a given backend.
for attr_name in dir(_lowercase ):
snake_case_ :int = None
if _re_tf_models.match(_lowercase ) is not None:
snake_case_ :int = tf_models
snake_case_ :List[str] = _re_tf_models.match(_lowercase ).groups()[0]
elif _re_flax_models.match(_lowercase ) is not None:
snake_case_ :List[Any] = flax_models
snake_case_ :Any = _re_flax_models.match(_lowercase ).groups()[0]
elif _re_pt_models.match(_lowercase ) is not None:
snake_case_ :Optional[Any] = pt_models
snake_case_ :int = _re_pt_models.match(_lowercase ).groups()[0]
if lookup_dict is not None:
while len(_lowercase ) > 0:
if attr_name in model_prefix_to_model_type:
snake_case_ :Optional[int] = True
break
# Try again after removing the last word in the name
snake_case_ :Optional[Any] = """""".join(camel_case_split(_lowercase )[:-1] )
snake_case_ :Optional[int] = set(list(pt_models.keys() ) + list(tf_models.keys() ) + list(flax_models.keys() ) )
snake_case_ :Optional[Any] = list(_lowercase )
all_models.sort()
snake_case_ :Optional[int] = {"""model_type""": all_models}
snake_case_ :Optional[int] = [pt_models[t] for t in all_models]
snake_case_ :Any = [tf_models[t] for t in all_models]
snake_case_ :Dict = [flax_models[t] for t in all_models]
# Now let's use the auto-mapping names to make sure
snake_case_ :Dict = {}
for t in all_models:
if t in transformers_module.models.auto.processing_auto.PROCESSOR_MAPPING_NAMES:
snake_case_ :Optional[Any] = """AutoProcessor"""
elif t in transformers_module.models.auto.tokenization_auto.TOKENIZER_MAPPING_NAMES:
snake_case_ :Tuple = """AutoTokenizer"""
elif t in transformers_module.models.auto.feature_extraction_auto.FEATURE_EXTRACTOR_MAPPING_NAMES:
snake_case_ :Tuple = """AutoFeatureExtractor"""
else:
# Default to AutoTokenizer if a model has nothing, for backward compatibility.
snake_case_ :str = """AutoTokenizer"""
snake_case_ :int = [processors[t] for t in all_models]
return pd.DataFrame(_lowercase )
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :List[Any] = [
transformers_module.models.auto.modeling_auto,
transformers_module.models.auto.modeling_tf_auto,
transformers_module.models.auto.modeling_flax_auto,
]
for pipeline_tag, model_mapping, auto_class in PIPELINE_TAGS_AND_AUTO_MODELS:
snake_case_ :Optional[int] = [model_mapping, f"""TF_{model_mapping}""", f"""FLAX_{model_mapping}"""]
snake_case_ :List[str] = [auto_class, f"""TF_{auto_class}""", f"""Flax_{auto_class}"""]
# Loop through all three frameworks
for module, cls, mapping in zip(_lowercase, _lowercase, _lowercase ):
# The type of pipeline may not exist in this framework
if not hasattr(_lowercase, _lowercase ):
continue
# First extract all model_names
snake_case_ :Tuple = []
for name in getattr(_lowercase, _lowercase ).values():
if isinstance(_lowercase, _lowercase ):
model_names.append(_lowercase )
else:
model_names.extend(list(_lowercase ) )
# Add pipeline tag and auto model class for those models
table.update({model_name: (pipeline_tag, cls) for model_name in model_names} )
return table
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :List[Any] = get_frameworks_table()
snake_case_ :str = Dataset.from_pandas(_lowercase )
snake_case_ :List[Any] = hf_hub_download(
"""huggingface/transformers-metadata""", """pipeline_tags.json""", repo_type="""dataset""", token=_lowercase )
snake_case_ :List[str] = Dataset.from_json(_lowercase )
snake_case_ :int = {
tags_dataset[i]["""model_class"""]: (tags_dataset[i]["""pipeline_tag"""], tags_dataset[i]["""auto_class"""])
for i in range(len(_lowercase ) )
}
snake_case_ :Optional[int] = update_pipeline_and_auto_class_table(_lowercase )
# Sort the model classes to avoid some nondeterministic updates to create false update commits.
snake_case_ :Tuple = sorted(table.keys() )
snake_case_ :Tuple = pd.DataFrame(
{
"""model_class""": model_classes,
"""pipeline_tag""": [table[m][0] for m in model_classes],
"""auto_class""": [table[m][1] for m in model_classes],
} )
snake_case_ :Union[str, Any] = Dataset.from_pandas(_lowercase )
with tempfile.TemporaryDirectory() as tmp_dir:
frameworks_dataset.to_json(os.path.join(_lowercase, """frameworks.json""" ) )
tags_dataset.to_json(os.path.join(_lowercase, """pipeline_tags.json""" ) )
if commit_sha is not None:
snake_case_ :Union[str, Any] = (
f"""Update with commit {commit_sha}\n\nSee: """
f"""https://github.com/huggingface/transformers/commit/{commit_sha}"""
)
else:
snake_case_ :List[Any] = """Update"""
upload_folder(
repo_id="""huggingface/transformers-metadata""", folder_path=_lowercase, repo_type="""dataset""", token=_lowercase, commit_message=_lowercase, )
def A_ ( ):
'''simple docstring'''
snake_case_ :List[Any] = {tag: cls for tag, _, cls in PIPELINE_TAGS_AND_AUTO_MODELS}
snake_case_ :Dict = transformers_module.pipelines.SUPPORTED_TASKS
snake_case_ :List[str] = []
for key in pipeline_tasks:
if key not in in_table:
snake_case_ :int = pipeline_tasks[key]["""pt"""]
if isinstance(_lowercase, (list, tuple) ):
snake_case_ :Any = model[0]
snake_case_ :str = model.__name__
if model not in in_table.values():
missing.append(_lowercase )
if len(_lowercase ) > 0:
snake_case_ :Optional[int] = """, """.join(_lowercase )
raise ValueError(
"""The following pipeline tags are not present in the `PIPELINE_TAGS_AND_AUTO_MODELS` constant inside """
f"""`utils/update_metadata.py`: {msg}. Please add them!""" )
if __name__ == "__main__":
__a = argparse.ArgumentParser()
parser.add_argument("--token", type=str, help="The token to use to push to the transformers-metadata dataset.")
parser.add_argument("--commit_sha", type=str, help="The sha of the commit going with this update.")
parser.add_argument("--check-only", action="store_true", help="Activate to just check all pipelines are present.")
__a = parser.parse_args()
if args.check_only:
check_pipeline_tags()
else:
update_metadata(args.token, args.commit_sha)
| 66 | 1 |
"""simple docstring"""
import os
import unittest
from huggingface_hub.utils import are_progress_bars_disabled
import transformers.models.bart.tokenization_bart
from transformers import logging
from transformers.testing_utils import CaptureLogger, mockenv, mockenv_context
from transformers.utils.logging import disable_progress_bar, enable_progress_bar
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: List[Any] ) -> Union[str, Any]:
snake_case_ :str = logging.get_logger()
# the current default level is logging.WARNING
snake_case_ :int = logging.get_verbosity()
logging.set_verbosity_error()
self.assertEqual(logger.getEffectiveLevel() , logging.get_verbosity() )
logging.set_verbosity_warning()
self.assertEqual(logger.getEffectiveLevel() , logging.get_verbosity() )
logging.set_verbosity_info()
self.assertEqual(logger.getEffectiveLevel() , logging.get_verbosity() )
logging.set_verbosity_debug()
self.assertEqual(logger.getEffectiveLevel() , logging.get_verbosity() )
# restore to the original level
logging.set_verbosity(snake_case )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Union[str, Any]:
snake_case_ :Any = logging.get_verbosity()
snake_case_ :Union[str, Any] = logging.get_logger("""transformers.models.bart.tokenization_bart""" )
snake_case_ :List[Any] = """Testing 1, 2, 3"""
# should be able to log warnings (if default settings weren't overridden by `pytest --log-level-all`)
if level_origin <= logging.WARNING:
with CaptureLogger(snake_case ) as cl:
logger.warning(snake_case )
self.assertEqual(cl.out , msg + """\n""" )
# this is setting the level for all of `transformers.*` loggers
logging.set_verbosity_error()
# should not be able to log warnings
with CaptureLogger(snake_case ) as cl:
logger.warning(snake_case )
self.assertEqual(cl.out , """""" )
# should be able to log warnings again
logging.set_verbosity_warning()
with CaptureLogger(snake_case ) as cl:
logger.warning(snake_case )
self.assertEqual(cl.out , msg + """\n""" )
# restore to the original level
logging.set_verbosity(snake_case )
@mockenv(TRANSFORMERS_VERBOSITY="""error""" )
def lowerCAmelCase_ ( self: List[str] ) -> Any:
# reset for the env var to take effect, next time some logger call is made
transformers.utils.logging._reset_library_root_logger()
# this action activates the env var
snake_case_ :Any = logging.get_logger("""transformers.models.bart.tokenization_bart""" )
snake_case_ :List[Any] = os.getenv("""TRANSFORMERS_VERBOSITY""" , snake_case )
snake_case_ :int = logging.log_levels[env_level_str]
snake_case_ :List[Any] = logging.get_verbosity()
self.assertEqual(
snake_case , snake_case , f"""TRANSFORMERS_VERBOSITY={env_level_str}/{env_level}, but internal verbosity is {current_level}""" , )
# restore to the original level
snake_case_ :Optional[int] = """"""
transformers.utils.logging._reset_library_root_logger()
@mockenv(TRANSFORMERS_VERBOSITY="""super-error""" )
def lowerCAmelCase_ ( self: Dict ) -> Optional[int]:
# reset for the env var to take effect, next time some logger call is made
transformers.utils.logging._reset_library_root_logger()
snake_case_ :Dict = logging.logging.getLogger()
with CaptureLogger(snake_case ) as cl:
# this action activates the env var
logging.get_logger("""transformers.models.bart.tokenization_bart""" )
self.assertIn("""Unknown option TRANSFORMERS_VERBOSITY=super-error""" , cl.out )
# no need to restore as nothing was changed
def lowerCAmelCase_ ( self: Any ) -> Tuple:
# testing `logger.warning_advice()`
transformers.utils.logging._reset_library_root_logger()
snake_case_ :List[Any] = logging.get_logger("""transformers.models.bart.tokenization_bart""" )
snake_case_ :Tuple = """Testing 1, 2, 3"""
with mockenv_context(TRANSFORMERS_NO_ADVISORY_WARNINGS="""1""" ):
# nothing should be logged as env var disables this method
with CaptureLogger(snake_case ) as cl:
logger.warning_advice(snake_case )
self.assertEqual(cl.out , """""" )
with mockenv_context(TRANSFORMERS_NO_ADVISORY_WARNINGS="""""" ):
# should log normally as TRANSFORMERS_NO_ADVISORY_WARNINGS is unset
with CaptureLogger(snake_case ) as cl:
logger.warning_advice(snake_case )
self.assertEqual(cl.out , msg + """\n""" )
def A_ ( ):
'''simple docstring'''
disable_progress_bar()
assert are_progress_bars_disabled()
enable_progress_bar()
assert not are_progress_bars_disabled()
| 66 |
"""simple docstring"""
import argparse
import glob
import logging
import os
from argparse import Namespace
from importlib import import_module
import numpy as np
import torch
from lightning_base import BaseTransformer, add_generic_args, generic_train
from seqeval.metrics import accuracy_score, fa_score, precision_score, recall_score
from torch.nn import CrossEntropyLoss
from torch.utils.data import DataLoader, TensorDataset
from utils_ner import TokenClassificationTask
__a = logging.getLogger(__name__)
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
_A : Union[str, Any] = """token-classification"""
def __init__( self: Any , snake_case: Tuple ) -> List[Any]:
if type(snake_case ) == dict:
snake_case_ :Optional[int] = Namespace(**snake_case )
snake_case_ :Optional[int] = import_module("""tasks""" )
try:
snake_case_ :Any = getattr(snake_case , hparams.task_type )
snake_case_ :TokenClassificationTask = token_classification_task_clazz()
except AttributeError:
raise ValueError(
f"""Task {hparams.task_type} needs to be defined as a TokenClassificationTask subclass in {module}. """
f"""Available tasks classes are: {TokenClassificationTask.__subclasses__()}""" )
snake_case_ :Any = self.token_classification_task.get_labels(hparams.labels )
snake_case_ :str = CrossEntropyLoss().ignore_index
super().__init__(snake_case , len(self.labels ) , self.mode )
def lowerCAmelCase_ ( self: Dict , **snake_case: List[Any] ) -> Any:
return self.model(**snake_case )
def lowerCAmelCase_ ( self: str , snake_case: Tuple , snake_case: List[Any] ) -> Optional[int]:
snake_case_ :List[str] = {"""input_ids""": batch[0], """attention_mask""": batch[1], """labels""": batch[3]}
if self.config.model_type != "distilbert":
snake_case_ :List[str] = (
batch[2] if self.config.model_type in ["""bert""", """xlnet"""] else None
) # XLM and RoBERTa don"t use token_type_ids
snake_case_ :Optional[Any] = self(**snake_case )
snake_case_ :List[str] = outputs[0]
# tensorboard_logs = {"loss": loss, "rate": self.lr_scheduler.get_last_lr()[-1]}
return {"loss": loss}
def lowerCAmelCase_ ( self: int ) -> Dict:
snake_case_ :List[Any] = self.hparams
for mode in ["train", "dev", "test"]:
snake_case_ :Optional[int] = self._feature_file(snake_case )
if os.path.exists(snake_case ) and not args.overwrite_cache:
logger.info("""Loading features from cached file %s""" , snake_case )
snake_case_ :Optional[int] = torch.load(snake_case )
else:
logger.info("""Creating features from dataset file at %s""" , args.data_dir )
snake_case_ :Optional[int] = self.token_classification_task.read_examples_from_file(args.data_dir , snake_case )
snake_case_ :Any = self.token_classification_task.convert_examples_to_features(
snake_case , self.labels , args.max_seq_length , self.tokenizer , cls_token_at_end=bool(self.config.model_type in ["""xlnet"""] ) , cls_token=self.tokenizer.cls_token , cls_token_segment_id=2 if self.config.model_type in ["""xlnet"""] else 0 , sep_token=self.tokenizer.sep_token , sep_token_extra=snake_case , pad_on_left=bool(self.config.model_type in ["""xlnet"""] ) , pad_token=self.tokenizer.pad_token_id , pad_token_segment_id=self.tokenizer.pad_token_type_id , pad_token_label_id=self.pad_token_label_id , )
logger.info("""Saving features into cached file %s""" , snake_case )
torch.save(snake_case , snake_case )
def lowerCAmelCase_ ( self: Optional[int] , snake_case: int , snake_case: int , snake_case: bool = False ) -> DataLoader:
snake_case_ :int = self._feature_file(snake_case )
logger.info("""Loading features from cached file %s""" , snake_case )
snake_case_ :str = torch.load(snake_case )
snake_case_ :Dict = torch.tensor([f.input_ids for f in features] , dtype=torch.long )
snake_case_ :str = torch.tensor([f.attention_mask for f in features] , dtype=torch.long )
if features[0].token_type_ids is not None:
snake_case_ :List[Any] = torch.tensor([f.token_type_ids for f in features] , dtype=torch.long )
else:
snake_case_ :List[str] = torch.tensor([0 for f in features] , dtype=torch.long )
# HACK(we will not use this anymore soon)
snake_case_ :Any = torch.tensor([f.label_ids for f in features] , dtype=torch.long )
return DataLoader(
TensorDataset(snake_case , snake_case , snake_case , snake_case ) , batch_size=snake_case )
def lowerCAmelCase_ ( self: List[str] , snake_case: Dict , snake_case: Union[str, Any] ) -> List[str]:
"""Compute validation""" ""
snake_case_ :List[str] = {"""input_ids""": batch[0], """attention_mask""": batch[1], """labels""": batch[3]}
if self.config.model_type != "distilbert":
snake_case_ :Dict = (
batch[2] if self.config.model_type in ["""bert""", """xlnet"""] else None
) # XLM and RoBERTa don"t use token_type_ids
snake_case_ :Dict = self(**snake_case )
snake_case_, snake_case_ :Dict = outputs[:2]
snake_case_ :Union[str, Any] = logits.detach().cpu().numpy()
snake_case_ :List[Any] = inputs["""labels"""].detach().cpu().numpy()
return {"val_loss": tmp_eval_loss.detach().cpu(), "pred": preds, "target": out_label_ids}
def lowerCAmelCase_ ( self: List[Any] , snake_case: int ) -> Tuple:
snake_case_ :Union[str, Any] = torch.stack([x["""val_loss"""] for x in outputs] ).mean()
snake_case_ :Tuple = np.concatenate([x["""pred"""] for x in outputs] , axis=0 )
snake_case_ :Tuple = np.argmax(snake_case , axis=2 )
snake_case_ :List[str] = np.concatenate([x["""target"""] for x in outputs] , axis=0 )
snake_case_ :Optional[Any] = dict(enumerate(self.labels ) )
snake_case_ :Dict = [[] for _ in range(out_label_ids.shape[0] )]
snake_case_ :Dict = [[] for _ in range(out_label_ids.shape[0] )]
for i in range(out_label_ids.shape[0] ):
for j in range(out_label_ids.shape[1] ):
if out_label_ids[i, j] != self.pad_token_label_id:
out_label_list[i].append(label_map[out_label_ids[i][j]] )
preds_list[i].append(label_map[preds[i][j]] )
snake_case_ :str = {
"""val_loss""": val_loss_mean,
"""accuracy_score""": accuracy_score(snake_case , snake_case ),
"""precision""": precision_score(snake_case , snake_case ),
"""recall""": recall_score(snake_case , snake_case ),
"""f1""": fa_score(snake_case , snake_case ),
}
snake_case_ :List[Any] = dict(results.items() )
snake_case_ :Union[str, Any] = results
return ret, preds_list, out_label_list
def lowerCAmelCase_ ( self: Optional[Any] , snake_case: Dict ) -> Optional[Any]:
# when stable
snake_case_, snake_case_, snake_case_ :Tuple = self._eval_end(snake_case )
snake_case_ :str = ret["""log"""]
return {"val_loss": logs["val_loss"], "log": logs, "progress_bar": logs}
def lowerCAmelCase_ ( self: Tuple , snake_case: Optional[int] ) -> Any:
# updating to test_epoch_end instead of deprecated test_end
snake_case_, snake_case_, snake_case_ :Any = self._eval_end(snake_case )
# Converting to the dict required by pl
# https://github.com/PyTorchLightning/pytorch-lightning/blob/master/\
# pytorch_lightning/trainer/logging.py#L139
snake_case_ :Optional[int] = ret["""log"""]
# `val_loss` is the key returned by `self._eval_end()` but actually refers to `test_loss`
return {"avg_test_loss": logs["val_loss"], "log": logs, "progress_bar": logs}
@staticmethod
def lowerCAmelCase_ ( snake_case: Any , snake_case: int ) -> Dict:
# Add NER specific options
BaseTransformer.add_model_specific_args(snake_case , snake_case )
parser.add_argument(
"""--task_type""" , default="""NER""" , type=snake_case , help="""Task type to fine tune in training (e.g. NER, POS, etc)""" )
parser.add_argument(
"""--max_seq_length""" , default=128 , type=snake_case , help=(
"""The maximum total input sequence length after tokenization. Sequences longer """
"""than this will be truncated, sequences shorter will be padded."""
) , )
parser.add_argument(
"""--labels""" , default="""""" , type=snake_case , help="""Path to a file containing all labels. If not specified, CoNLL-2003 labels are used.""" , )
parser.add_argument(
"""--gpus""" , default=0 , type=snake_case , help="""The number of GPUs allocated for this, it is by default 0 meaning none""" , )
parser.add_argument(
"""--overwrite_cache""" , action="""store_true""" , help="""Overwrite the cached training and evaluation sets""" )
return parser
if __name__ == "__main__":
__a = argparse.ArgumentParser()
add_generic_args(parser, os.getcwd())
__a = NERTransformer.add_model_specific_args(parser, os.getcwd())
__a = parser.parse_args()
__a = NERTransformer(args)
__a = generic_train(model, args)
if args.do_predict:
# See https://github.com/huggingface/transformers/issues/3159
# pl use this default format to create a checkpoint:
# https://github.com/PyTorchLightning/pytorch-lightning/blob/master\
# /pytorch_lightning/callbacks/model_checkpoint.py#L322
__a = sorted(glob.glob(os.path.join(args.output_dir, "checkpoint-epoch=*.ckpt"), recursive=True))
__a = model.load_from_checkpoint(checkpoints[-1])
trainer.test(model)
| 66 | 1 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available, is_vision_available
__a = {
"configuration_mask2former": [
"MASK2FORMER_PRETRAINED_CONFIG_ARCHIVE_MAP",
"Mask2FormerConfig",
],
}
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["Mask2FormerImageProcessor"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"MASK2FORMER_PRETRAINED_MODEL_ARCHIVE_LIST",
"Mask2FormerForUniversalSegmentation",
"Mask2FormerModel",
"Mask2FormerPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_maskaformer import MASK2FORMER_PRETRAINED_CONFIG_ARCHIVE_MAP, MaskaFormerConfig
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .image_processing_maskaformer import MaskaFormerImageProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_maskaformer import (
MASK2FORMER_PRETRAINED_MODEL_ARCHIVE_LIST,
MaskaFormerForUniversalSegmentation,
MaskaFormerModel,
MaskaFormerPreTrainedModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure)
| 66 |
"""simple docstring"""
from math import factorial
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Optional[int] , snake_case: Dict , snake_case: int ) -> Tuple:
snake_case_ :List[Any] = real
if isinstance(snake_case , snake_case ):
snake_case_ :Tuple = [1] * rank
else:
snake_case_ :Optional[Any] = rank
def __repr__( self: List[str] ) -> Tuple:
return (
f"""{self.real}+"""
f"""{'+'.join(str(snake_case )+'E'+str(n+1 )for n,dual in enumerate(self.duals ) )}"""
)
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
snake_case_ :Any = self.duals.copy()
while cur[-1] == 0:
cur.pop(-1 )
return Dual(self.real , snake_case )
def __add__( self: Optional[int] , snake_case: Dict ) -> List[str]:
if not isinstance(snake_case , snake_case ):
return Dual(self.real + other , self.duals )
snake_case_ :List[Any] = self.duals.copy()
snake_case_ :Tuple = other.duals.copy()
if len(snake_case ) > len(snake_case ):
o_dual.extend([1] * (len(snake_case ) - len(snake_case )) )
elif len(snake_case ) < len(snake_case ):
s_dual.extend([1] * (len(snake_case ) - len(snake_case )) )
snake_case_ :Dict = []
for i in range(len(snake_case ) ):
new_duals.append(s_dual[i] + o_dual[i] )
return Dual(self.real + other.real , snake_case )
_A : str = __add__
def __sub__( self: Tuple , snake_case: Union[str, Any] ) -> Tuple:
return self + other * -1
def __mul__( self: str , snake_case: Tuple ) -> Optional[Any]:
if not isinstance(snake_case , snake_case ):
snake_case_ :Dict = []
for i in self.duals:
new_duals.append(i * other )
return Dual(self.real * other , snake_case )
snake_case_ :int = [0] * (len(self.duals ) + len(other.duals ) + 1)
for i, item in enumerate(self.duals ):
for j, jtem in enumerate(other.duals ):
new_duals[i + j + 1] += item * jtem
for k in range(len(self.duals ) ):
new_duals[k] += self.duals[k] * other.real
for index in range(len(other.duals ) ):
new_duals[index] += other.duals[index] * self.real
return Dual(self.real * other.real , snake_case )
_A : int = __mul__
def __truediv__( self: List[str] , snake_case: List[str] ) -> List[str]:
if not isinstance(snake_case , snake_case ):
snake_case_ :Optional[Any] = []
for i in self.duals:
new_duals.append(i / other )
return Dual(self.real / other , snake_case )
raise ValueError
def __floordiv__( self: int , snake_case: List[Any] ) -> Any:
if not isinstance(snake_case , snake_case ):
snake_case_ :Optional[int] = []
for i in self.duals:
new_duals.append(i // other )
return Dual(self.real // other , snake_case )
raise ValueError
def __pow__( self: Optional[Any] , snake_case: Optional[int] ) -> List[Any]:
if n < 0 or isinstance(snake_case , snake_case ):
raise ValueError("""power must be a positive integer""" )
if n == 0:
return 1
if n == 1:
return self
snake_case_ :str = self
for _ in range(n - 1 ):
x *= self
return x
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
if not callable(_lowercase ):
raise ValueError("""differentiate() requires a function as input for func""" )
if not isinstance(_lowercase, (float, int) ):
raise ValueError("""differentiate() requires a float as input for position""" )
if not isinstance(_lowercase, _lowercase ):
raise ValueError("""differentiate() requires an int as input for order""" )
snake_case_ :Optional[Any] = Dual(_lowercase, 1 )
snake_case_ :List[Any] = func(_lowercase )
if order == 0:
return result.real
return result.duals[order - 1] * factorial(_lowercase )
if __name__ == "__main__":
import doctest
doctest.testmod()
def A_ ( _lowercase ):
'''simple docstring'''
return y**2 * y**4
print(differentiate(f, 9, 2))
| 66 | 1 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import _LazyModule
__a = {"tokenization_byt5": ["ByT5Tokenizer"]}
if TYPE_CHECKING:
from .tokenization_byta import ByTaTokenizer
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 |
"""simple docstring"""
from __future__ import annotations
__a = 10
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Union[str, Any] = 1
snake_case_ :List[str] = max(_lowercase )
while placement <= max_digit:
# declare and initialize empty buckets
snake_case_ :list[list] = [[] for _ in range(_lowercase )]
# split list_of_ints between the buckets
for i in list_of_ints:
snake_case_ :Any = int((i / placement) % RADIX )
buckets[tmp].append(_lowercase )
# put each buckets' contents into list_of_ints
snake_case_ :Optional[Any] = 0
for b in range(_lowercase ):
for i in buckets[b]:
snake_case_ :Union[str, Any] = i
a += 1
# move to next
placement *= RADIX
return list_of_ints
if __name__ == "__main__":
import doctest
doctest.testmod()
| 66 | 1 |
"""simple docstring"""
from ...configuration_utils import PretrainedConfig
from ...utils import logging
__a = logging.get_logger(__name__)
__a = {
"google/switch-base-8": "https://huggingface.co/google/switch-base-8/blob/main/config.json",
}
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
_A : Tuple = """switch_transformers"""
_A : Tuple = ["""past_key_values"""]
_A : Tuple = {"""hidden_size""": """d_model""", """num_attention_heads""": """num_heads""", """num_hidden_layers""": """num_layers"""}
def __init__( self: Union[str, Any] , snake_case: Optional[Any]=32_128 , snake_case: List[Any]=768 , snake_case: List[str]=64 , snake_case: str=2_048 , snake_case: Optional[int]=64 , snake_case: str=12 , snake_case: int=3 , snake_case: Tuple=12 , snake_case: Dict=3 , snake_case: Dict=12 , snake_case: Tuple=8 , snake_case: int=False , snake_case: Tuple=0.0_1 , snake_case: int="float32" , snake_case: Optional[Any]=False , snake_case: Union[str, Any]=32 , snake_case: str=128 , snake_case: Tuple=0.1 , snake_case: Optional[int]=1E-6 , snake_case: List[str]=0.0_0_1 , snake_case: List[str]=0.0_0_1 , snake_case: List[str]=1.0 , snake_case: Dict="relu" , snake_case: str=True , snake_case: Union[str, Any]=False , snake_case: Any=True , snake_case: List[Any]=0 , snake_case: Tuple=1 , **snake_case: int , ) -> Dict:
snake_case_ :List[str] = vocab_size
snake_case_ :List[str] = d_model
snake_case_ :Optional[Any] = d_kv
snake_case_ :Union[str, Any] = d_ff
snake_case_ :Union[str, Any] = num_sparse_encoder_layers
snake_case_ :Any = num_layers
snake_case_ :Optional[Any] = (
num_decoder_layers if num_decoder_layers is not None else self.num_layers
) # default = symmetry
snake_case_ :List[Any] = num_sparse_decoder_layers
# This tells us, each how many encoder layer we'll have to set a sparse layer.
if self.num_sparse_encoder_layers > 0:
snake_case_ :Optional[int] = self.num_layers // self.num_sparse_encoder_layers
else:
snake_case_ :int = self.num_layers # HACK: this will create 0 sparse layers
# This tells us, each how many encoder layer we'll have to set a sparse layer.
if self.num_sparse_decoder_layers > 0:
snake_case_ :Tuple = self.num_decoder_layers // self.num_sparse_decoder_layers
else:
snake_case_ :Any = self.num_decoder_layers # HACK: this will create 0 sparse layers
snake_case_ :int = num_heads
snake_case_ :Any = num_experts
snake_case_ :Optional[Any] = expert_capacity
snake_case_ :int = router_bias
snake_case_ :Optional[Any] = router_jitter_noise
if router_dtype not in ["float32", "float16", "bfloat16"]:
raise ValueError(f"""`router_dtype` must be one of 'float32', 'float16' or 'bfloat16', got {router_dtype}""" )
snake_case_ :str = router_dtype
snake_case_ :Optional[Any] = router_ignore_padding_tokens
snake_case_ :str = relative_attention_num_buckets
snake_case_ :Tuple = relative_attention_max_distance
snake_case_ :List[str] = dropout_rate
snake_case_ :List[str] = layer_norm_epsilon
snake_case_ :Dict = initializer_factor
snake_case_ :Tuple = feed_forward_proj
snake_case_ :List[str] = use_cache
snake_case_ :List[str] = add_router_probs
snake_case_ :Optional[Any] = router_z_loss_coef
snake_case_ :List[str] = router_aux_loss_coef
snake_case_ :Dict = self.feed_forward_proj.split("""-""" )
snake_case_ :int = act_info[-1]
snake_case_ :Union[str, Any] = act_info[0] == """gated"""
if len(snake_case ) > 1 and act_info[0] != "gated" or len(snake_case ) > 2:
raise ValueError(
f"""`feed_forward_proj`: {feed_forward_proj} is not a valid activation function of the dense layer."""
"""Please make sure `feed_forward_proj` is of the format `gated-{ACT_FN}` or `{ACT_FN}`, e.g. """
"""'gated-gelu' or 'relu'""" )
# for backwards compatibility
if feed_forward_proj == "gated-gelu":
snake_case_ :Tuple = """gelu_new"""
super().__init__(
pad_token_id=snake_case , eos_token_id=snake_case , is_encoder_decoder=snake_case , **snake_case , )
| 66 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_sentencepiece_available,
is_tokenizers_available,
is_torch_available,
)
__a = {"configuration_reformer": ["REFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP", "ReformerConfig"]}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["ReformerTokenizer"]
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["ReformerTokenizerFast"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"REFORMER_PRETRAINED_MODEL_ARCHIVE_LIST",
"ReformerAttention",
"ReformerForMaskedLM",
"ReformerForQuestionAnswering",
"ReformerForSequenceClassification",
"ReformerLayer",
"ReformerModel",
"ReformerModelWithLMHead",
"ReformerPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_reformer import REFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP, ReformerConfig
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_reformer import ReformerTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_reformer_fast import ReformerTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_reformer import (
REFORMER_PRETRAINED_MODEL_ARCHIVE_LIST,
ReformerAttention,
ReformerForMaskedLM,
ReformerForQuestionAnswering,
ReformerForSequenceClassification,
ReformerLayer,
ReformerModel,
ReformerModelWithLMHead,
ReformerPreTrainedModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 | 1 |
"""simple docstring"""
from ..utils import DummyObject, requires_backends
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Union[str, Any] = ["""sentencepiece"""]
def __init__( self: Optional[Any] , *snake_case: int , **snake_case: List[Any] ) -> Optional[Any]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : str = ["""sentencepiece"""]
def __init__( self: List[Any] , *snake_case: Optional[Any] , **snake_case: List[Any] ) -> Any:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : int = ["""sentencepiece"""]
def __init__( self: List[str] , *snake_case: Dict , **snake_case: int ) -> Union[str, Any]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Any = ["""sentencepiece"""]
def __init__( self: int , *snake_case: int , **snake_case: List[str] ) -> str:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Dict = ["""sentencepiece"""]
def __init__( self: List[str] , *snake_case: List[str] , **snake_case: Dict ) -> Optional[Any]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Any = ["""sentencepiece"""]
def __init__( self: Dict , *snake_case: str , **snake_case: Dict ) -> Any:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Optional[Any] = ["""sentencepiece"""]
def __init__( self: Tuple , *snake_case: List[Any] , **snake_case: Optional[Any] ) -> Union[str, Any]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Dict = ["""sentencepiece"""]
def __init__( self: Union[str, Any] , *snake_case: List[str] , **snake_case: Optional[int] ) -> Optional[int]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Tuple = ["""sentencepiece"""]
def __init__( self: Optional[Any] , *snake_case: List[str] , **snake_case: Tuple ) -> Optional[Any]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : str = ["""sentencepiece"""]
def __init__( self: List[Any] , *snake_case: Optional[int] , **snake_case: int ) -> str:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Optional[int] = ["""sentencepiece"""]
def __init__( self: Union[str, Any] , *snake_case: str , **snake_case: Union[str, Any] ) -> List[Any]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : int = ["""sentencepiece"""]
def __init__( self: List[Any] , *snake_case: Optional[Any] , **snake_case: int ) -> List[str]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : str = ["""sentencepiece"""]
def __init__( self: List[str] , *snake_case: int , **snake_case: Tuple ) -> Optional[Any]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Optional[int] = ["""sentencepiece"""]
def __init__( self: str , *snake_case: Optional[int] , **snake_case: Any ) -> Union[str, Any]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : List[str] = ["""sentencepiece"""]
def __init__( self: str , *snake_case: Tuple , **snake_case: int ) -> str:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Dict = ["""sentencepiece"""]
def __init__( self: Dict , *snake_case: int , **snake_case: List[str] ) -> List[str]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Union[str, Any] = ["""sentencepiece"""]
def __init__( self: List[Any] , *snake_case: List[str] , **snake_case: Optional[int] ) -> Union[str, Any]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Dict = ["""sentencepiece"""]
def __init__( self: Optional[int] , *snake_case: str , **snake_case: List[Any] ) -> List[Any]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Dict = ["""sentencepiece"""]
def __init__( self: Tuple , *snake_case: str , **snake_case: Tuple ) -> int:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : str = ["""sentencepiece"""]
def __init__( self: Union[str, Any] , *snake_case: Optional[Any] , **snake_case: Optional[Any] ) -> Optional[Any]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Any = ["""sentencepiece"""]
def __init__( self: Tuple , *snake_case: Optional[Any] , **snake_case: Any ) -> Optional[Any]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Union[str, Any] = ["""sentencepiece"""]
def __init__( self: Any , *snake_case: Optional[Any] , **snake_case: List[str] ) -> str:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : int = ["""sentencepiece"""]
def __init__( self: str , *snake_case: List[Any] , **snake_case: Any ) -> int:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Dict = ["""sentencepiece"""]
def __init__( self: Optional[int] , *snake_case: Optional[int] , **snake_case: Union[str, Any] ) -> Optional[Any]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Union[str, Any] = ["""sentencepiece"""]
def __init__( self: Optional[int] , *snake_case: str , **snake_case: Dict ) -> str:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Union[str, Any] = ["""sentencepiece"""]
def __init__( self: Union[str, Any] , *snake_case: int , **snake_case: List[Any] ) -> List[str]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Any = ["""sentencepiece"""]
def __init__( self: List[Any] , *snake_case: str , **snake_case: Optional[Any] ) -> Union[str, Any]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : str = ["""sentencepiece"""]
def __init__( self: Union[str, Any] , *snake_case: List[str] , **snake_case: Union[str, Any] ) -> Tuple:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : Optional[int] = ["""sentencepiece"""]
def __init__( self: List[str] , *snake_case: str , **snake_case: Any ) -> Optional[int]:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : int = ["""sentencepiece"""]
def __init__( self: Tuple , *snake_case: Any , **snake_case: Optional[Any] ) -> Dict:
requires_backends(self , ["""sentencepiece"""] )
class lowerCamelCase ( metaclass=_lowerCAmelCase ):
'''simple docstring'''
_A : List[str] = ["""sentencepiece"""]
def __init__( self: str , *snake_case: List[str] , **snake_case: Optional[int] ) -> Tuple:
requires_backends(self , ["""sentencepiece"""] )
| 66 |
"""simple docstring"""
import gc
import unittest
from diffusers import FlaxControlNetModel, FlaxStableDiffusionControlNetPipeline
from diffusers.utils import is_flax_available, load_image, slow
from diffusers.utils.testing_utils import require_flax
if is_flax_available():
import jax
import jax.numpy as jnp
from flax.jax_utils import replicate
from flax.training.common_utils import shard
@slow
@require_flax
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: List[Any] ) -> Any:
# clean up the VRAM after each test
super().tearDown()
gc.collect()
def lowerCAmelCase_ ( self: Tuple ) -> Any:
snake_case_, snake_case_ :List[str] = FlaxControlNetModel.from_pretrained(
"""lllyasviel/sd-controlnet-canny""" , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_, snake_case_ :Union[str, Any] = FlaxStableDiffusionControlNetPipeline.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" , controlnet=snake_case , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_ :Union[str, Any] = controlnet_params
snake_case_ :Union[str, Any] = """bird"""
snake_case_ :List[Any] = jax.device_count()
snake_case_ :List[Any] = pipe.prepare_text_inputs([prompts] * num_samples )
snake_case_ :List[str] = load_image(
"""https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/sd_controlnet/bird_canny.png""" )
snake_case_ :List[str] = pipe.prepare_image_inputs([canny_image] * num_samples )
snake_case_ :Any = jax.random.PRNGKey(0 )
snake_case_ :List[str] = jax.random.split(snake_case , jax.device_count() )
snake_case_ :List[Any] = replicate(snake_case )
snake_case_ :List[str] = shard(snake_case )
snake_case_ :str = shard(snake_case )
snake_case_ :Dict = pipe(
prompt_ids=snake_case , image=snake_case , params=snake_case , prng_seed=snake_case , num_inference_steps=50 , jit=snake_case , ).images
assert images.shape == (jax.device_count(), 1, 768, 512, 3)
snake_case_ :str = images.reshape((images.shape[0] * images.shape[1],) + images.shape[-3:] )
snake_case_ :Union[str, Any] = images[0, 253:256, 253:256, -1]
snake_case_ :str = jnp.asarray(jax.device_get(image_slice.flatten() ) )
snake_case_ :Dict = jnp.array(
[0.1_6_7_9_6_9, 0.1_1_6_6_9_9, 0.0_8_1_5_4_3, 0.1_5_4_2_9_7, 0.1_3_2_8_1_2, 0.1_0_8_8_8_7, 0.1_6_9_9_2_2, 0.1_6_9_9_2_2, 0.2_0_5_0_7_8] )
print(f"""output_slice: {output_slice}""" )
assert jnp.abs(output_slice - expected_slice ).max() < 1E-2
def lowerCAmelCase_ ( self: int ) -> Dict:
snake_case_, snake_case_ :List[Any] = FlaxControlNetModel.from_pretrained(
"""lllyasviel/sd-controlnet-openpose""" , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_, snake_case_ :int = FlaxStableDiffusionControlNetPipeline.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" , controlnet=snake_case , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_ :str = controlnet_params
snake_case_ :Optional[int] = """Chef in the kitchen"""
snake_case_ :Union[str, Any] = jax.device_count()
snake_case_ :Any = pipe.prepare_text_inputs([prompts] * num_samples )
snake_case_ :str = load_image(
"""https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/sd_controlnet/pose.png""" )
snake_case_ :Optional[Any] = pipe.prepare_image_inputs([pose_image] * num_samples )
snake_case_ :str = jax.random.PRNGKey(0 )
snake_case_ :str = jax.random.split(snake_case , jax.device_count() )
snake_case_ :Tuple = replicate(snake_case )
snake_case_ :str = shard(snake_case )
snake_case_ :int = shard(snake_case )
snake_case_ :List[str] = pipe(
prompt_ids=snake_case , image=snake_case , params=snake_case , prng_seed=snake_case , num_inference_steps=50 , jit=snake_case , ).images
assert images.shape == (jax.device_count(), 1, 768, 512, 3)
snake_case_ :str = images.reshape((images.shape[0] * images.shape[1],) + images.shape[-3:] )
snake_case_ :int = images[0, 253:256, 253:256, -1]
snake_case_ :Dict = jnp.asarray(jax.device_get(image_slice.flatten() ) )
snake_case_ :Optional[int] = jnp.array(
[[0.2_7_1_4_8_4, 0.2_6_1_7_1_9, 0.2_7_5_3_9_1, 0.2_7_7_3_4_4, 0.2_7_9_2_9_7, 0.2_9_1_0_1_6, 0.2_9_4_9_2_2, 0.3_0_2_7_3_4, 0.3_0_2_7_3_4]] )
print(f"""output_slice: {output_slice}""" )
assert jnp.abs(output_slice - expected_slice ).max() < 1E-2
| 66 | 1 |
"""simple docstring"""
from __future__ import annotations
def A_ ( _lowercase ):
'''simple docstring'''
if len(_lowercase ) == 0:
return []
snake_case_, snake_case_ :Tuple = min(_lowercase ), max(_lowercase )
snake_case_ :Tuple = int(max_value - min_value ) + 1
snake_case_ :list[list] = [[] for _ in range(_lowercase )]
for i in my_list:
buckets[int(i - min_value )].append(_lowercase )
return [v for bucket in buckets for v in sorted(_lowercase )]
if __name__ == "__main__":
from doctest import testmod
testmod()
assert bucket_sort([4, 5, 3, 2, 1]) == [1, 2, 3, 4, 5]
assert bucket_sort([0, 1, -10, 15, 2, -2]) == [-10, -2, 0, 1, 2, 15]
| 66 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
__a = {
"configuration_mobilebert": [
"MOBILEBERT_PRETRAINED_CONFIG_ARCHIVE_MAP",
"MobileBertConfig",
"MobileBertOnnxConfig",
],
"tokenization_mobilebert": ["MobileBertTokenizer"],
}
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["MobileBertTokenizerFast"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST",
"MobileBertForMaskedLM",
"MobileBertForMultipleChoice",
"MobileBertForNextSentencePrediction",
"MobileBertForPreTraining",
"MobileBertForQuestionAnswering",
"MobileBertForSequenceClassification",
"MobileBertForTokenClassification",
"MobileBertLayer",
"MobileBertModel",
"MobileBertPreTrainedModel",
"load_tf_weights_in_mobilebert",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"TF_MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST",
"TFMobileBertForMaskedLM",
"TFMobileBertForMultipleChoice",
"TFMobileBertForNextSentencePrediction",
"TFMobileBertForPreTraining",
"TFMobileBertForQuestionAnswering",
"TFMobileBertForSequenceClassification",
"TFMobileBertForTokenClassification",
"TFMobileBertMainLayer",
"TFMobileBertModel",
"TFMobileBertPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_mobilebert import (
MOBILEBERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
MobileBertConfig,
MobileBertOnnxConfig,
)
from .tokenization_mobilebert import MobileBertTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_mobilebert_fast import MobileBertTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_mobilebert import (
MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST,
MobileBertForMaskedLM,
MobileBertForMultipleChoice,
MobileBertForNextSentencePrediction,
MobileBertForPreTraining,
MobileBertForQuestionAnswering,
MobileBertForSequenceClassification,
MobileBertForTokenClassification,
MobileBertLayer,
MobileBertModel,
MobileBertPreTrainedModel,
load_tf_weights_in_mobilebert,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_mobilebert import (
TF_MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST,
TFMobileBertForMaskedLM,
TFMobileBertForMultipleChoice,
TFMobileBertForNextSentencePrediction,
TFMobileBertForPreTraining,
TFMobileBertForQuestionAnswering,
TFMobileBertForSequenceClassification,
TFMobileBertForTokenClassification,
TFMobileBertMainLayer,
TFMobileBertModel,
TFMobileBertPreTrainedModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 | 1 |
"""simple docstring"""
import argparse
import logging
import os
import re
import tensorflow as tf
from transformers import (
AutoConfig,
AutoTokenizer,
DataCollatorForLanguageModeling,
PushToHubCallback,
TFAutoModelForMaskedLM,
create_optimizer,
)
__a = logging.getLogger(__name__)
__a = tf.data.AUTOTUNE
def A_ ( ):
'''simple docstring'''
snake_case_ :Optional[int] = argparse.ArgumentParser(description="""Train a masked language model on TPU.""" )
parser.add_argument(
"""--pretrained_model_config""", type=_lowercase, default="""roberta-base""", help="""The model config to use. Note that we don't copy the model's weights, only the config!""", )
parser.add_argument(
"""--tokenizer""", type=_lowercase, default="""unigram-tokenizer-wikitext""", help="""The name of the tokenizer to load. We use the pretrained tokenizer to initialize the model's vocab size.""", )
parser.add_argument(
"""--per_replica_batch_size""", type=_lowercase, default=8, help="""Batch size per TPU core.""", )
parser.add_argument(
"""--no_tpu""", action="""store_true""", help="""If set, run on CPU and don't try to initialize a TPU. Useful for debugging on non-TPU instances.""", )
parser.add_argument(
"""--tpu_name""", type=_lowercase, help="""Name of TPU resource to initialize. Should be blank on Colab, and 'local' on TPU VMs.""", default="""local""", )
parser.add_argument(
"""--tpu_zone""", type=_lowercase, help="""Google cloud zone that TPU resource is located in. Only used for non-Colab TPU nodes.""", )
parser.add_argument(
"""--gcp_project""", type=_lowercase, help="""Google cloud project name. Only used for non-Colab TPU nodes.""" )
parser.add_argument(
"""--bfloat16""", action="""store_true""", help="""Use mixed-precision bfloat16 for training. This is the recommended lower-precision format for TPU.""", )
parser.add_argument(
"""--train_dataset""", type=_lowercase, help="""Path to training dataset to load. If the path begins with `gs://`"""
""" then the dataset will be loaded from a Google Cloud Storage bucket.""", )
parser.add_argument(
"""--shuffle_buffer_size""", type=_lowercase, default=2**18, help="""Size of the shuffle buffer (in samples)""", )
parser.add_argument(
"""--eval_dataset""", type=_lowercase, help="""Path to evaluation dataset to load. If the path begins with `gs://`"""
""" then the dataset will be loaded from a Google Cloud Storage bucket.""", )
parser.add_argument(
"""--num_epochs""", type=_lowercase, default=1, help="""Number of epochs to train for.""", )
parser.add_argument(
"""--learning_rate""", type=_lowercase, default=1e-4, help="""Learning rate to use for training.""", )
parser.add_argument(
"""--weight_decay_rate""", type=_lowercase, default=1e-3, help="""Weight decay rate to use for training.""", )
parser.add_argument(
"""--max_length""", type=_lowercase, default=512, help="""Maximum length of tokenized sequences. Should match the setting used in prepare_tfrecord_shards.py""", )
parser.add_argument(
"""--mlm_probability""", type=_lowercase, default=0.15, help="""Fraction of tokens to mask during training.""", )
parser.add_argument("""--output_dir""", type=_lowercase, required=_lowercase, help="""Path to save model checkpoints to.""" )
parser.add_argument("""--hub_model_id""", type=_lowercase, help="""Model ID to upload to on the Hugging Face Hub.""" )
snake_case_ :Dict = parser.parse_args()
return args
def A_ ( _lowercase ):
'''simple docstring'''
try:
if args.tpu_name:
snake_case_ :List[str] = tf.distribute.cluster_resolver.TPUClusterResolver(
args.tpu_name, zone=args.tpu_zone, project=args.gcp_project )
else:
snake_case_ :List[Any] = tf.distribute.cluster_resolver.TPUClusterResolver()
except ValueError:
raise RuntimeError(
"""Couldn't connect to TPU! Most likely you need to specify --tpu_name, --tpu_zone, or """
"""--gcp_project. When running on a TPU VM, use --tpu_name local.""" )
tf.config.experimental_connect_to_cluster(_lowercase )
tf.tpu.experimental.initialize_tpu_system(_lowercase )
return tpu
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Optional[int] = 0
for file in file_list:
snake_case_ :List[Any] = file.split("""/""" )[-1]
snake_case_ :Union[str, Any] = re.search(r"""-\d+-(\d+)\.tfrecord""", _lowercase ).group(1 )
snake_case_ :List[str] = int(_lowercase )
num_samples += sample_count
return num_samples
def A_ ( _lowercase, _lowercase, _lowercase, _lowercase, _lowercase, _lowercase=None ):
'''simple docstring'''
snake_case_ :Optional[int] = count_samples(_lowercase )
snake_case_ :Union[str, Any] = tf.data.Dataset.from_tensor_slices(_lowercase )
if shuffle:
snake_case_ :List[str] = dataset.shuffle(len(_lowercase ) )
snake_case_ :Any = tf.data.TFRecordDataset(_lowercase, num_parallel_reads=_lowercase )
# TF can't infer the total sample count because it doesn't read all the records yet, so we assert it here
snake_case_ :Optional[Any] = dataset.apply(tf.data.experimental.assert_cardinality(_lowercase ) )
snake_case_ :Any = dataset.map(_lowercase, num_parallel_calls=_lowercase )
if shuffle:
assert shuffle_buffer_size is not None
snake_case_ :Dict = dataset.shuffle(args.shuffle_buffer_size )
snake_case_ :Dict = dataset.batch(_lowercase, drop_remainder=_lowercase )
snake_case_ :Any = dataset.map(_lowercase, num_parallel_calls=_lowercase )
snake_case_ :Optional[int] = dataset.prefetch(_lowercase )
return dataset
def A_ ( _lowercase ):
'''simple docstring'''
if not args.no_tpu:
snake_case_ :Optional[Any] = initialize_tpu(_lowercase )
snake_case_ :Dict = tf.distribute.TPUStrategy(_lowercase )
else:
snake_case_ :Tuple = tf.distribute.OneDeviceStrategy(device="""/gpu:0""" )
if args.bfloataa:
tf.keras.mixed_precision.set_global_policy("""mixed_bfloat16""" )
snake_case_ :List[Any] = AutoTokenizer.from_pretrained(args.tokenizer )
snake_case_ :Any = AutoConfig.from_pretrained(args.pretrained_model_config )
snake_case_ :List[str] = tokenizer.vocab_size
snake_case_ :Optional[int] = tf.io.gfile.glob(os.path.join(args.train_dataset, """*.tfrecord""" ) )
if not training_records:
raise ValueError(f"""No .tfrecord files found in {args.train_dataset}.""" )
snake_case_ :Union[str, Any] = tf.io.gfile.glob(os.path.join(args.eval_dataset, """*.tfrecord""" ) )
if not eval_records:
raise ValueError(f"""No .tfrecord files found in {args.eval_dataset}.""" )
snake_case_ :Optional[Any] = count_samples(_lowercase )
snake_case_ :Union[str, Any] = num_train_samples // (args.per_replica_batch_size * strategy.num_replicas_in_sync)
snake_case_ :List[str] = steps_per_epoch * args.num_epochs
with strategy.scope():
snake_case_ :List[str] = TFAutoModelForMaskedLM.from_config(_lowercase )
model(model.dummy_inputs ) # Pass some dummy inputs through the model to ensure all the weights are built
snake_case_, snake_case_ :Optional[Any] = create_optimizer(
num_train_steps=_lowercase, num_warmup_steps=total_train_steps // 20, init_lr=args.learning_rate, weight_decay_rate=args.weight_decay_rate, )
# Transformers models compute the right loss for their task by default when labels are passed, and will
# use this for training unless you specify your own loss function in compile().
model.compile(optimizer=_lowercase, metrics=["""accuracy"""] )
def decode_fn(_lowercase ):
snake_case_ :Tuple = {
"""input_ids""": tf.io.FixedLenFeature(dtype=tf.intaa, shape=(args.max_length,) ),
"""attention_mask""": tf.io.FixedLenFeature(dtype=tf.intaa, shape=(args.max_length,) ),
}
return tf.io.parse_single_example(_lowercase, _lowercase )
# Many of the data collators in Transformers are TF-compilable when return_tensors == "tf", so we can
# use their methods in our data pipeline.
snake_case_ :Any = DataCollatorForLanguageModeling(
tokenizer=_lowercase, mlm_probability=args.mlm_probability, mlm=_lowercase, return_tensors="""tf""" )
def mask_with_collator(_lowercase ):
# TF really needs an isin() function
snake_case_ :List[str] = (
~tf.cast(batch["""attention_mask"""], tf.bool )
| (batch["""input_ids"""] == tokenizer.cls_token_id)
| (batch["""input_ids"""] == tokenizer.sep_token_id)
)
snake_case_, snake_case_ :Union[str, Any] = data_collator.tf_mask_tokens(
batch["""input_ids"""], vocab_size=len(_lowercase ), mask_token_id=tokenizer.mask_token_id, special_tokens_mask=_lowercase, )
return batch
snake_case_ :Optional[Any] = args.per_replica_batch_size * strategy.num_replicas_in_sync
snake_case_ :Union[str, Any] = prepare_dataset(
_lowercase, decode_fn=_lowercase, mask_fn=_lowercase, batch_size=_lowercase, shuffle=_lowercase, shuffle_buffer_size=args.shuffle_buffer_size, )
snake_case_ :Optional[Any] = prepare_dataset(
_lowercase, decode_fn=_lowercase, mask_fn=_lowercase, batch_size=_lowercase, shuffle=_lowercase, )
snake_case_ :List[Any] = []
if args.hub_model_id:
callbacks.append(
PushToHubCallback(output_dir=args.output_dir, hub_model_id=args.hub_model_id, tokenizer=_lowercase ) )
model.fit(
_lowercase, validation_data=_lowercase, epochs=args.num_epochs, callbacks=_lowercase, )
model.save_pretrained(args.output_dir )
if __name__ == "__main__":
__a = parse_args()
main(args)
| 66 |
"""simple docstring"""
import argparse
import json
import os
from collections import OrderedDict
import numpy as np
import tensorflow as tf
import torch
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Union[str, Any] = os.path.join(args.tf_model_dir, """parameters.json""" )
snake_case_ :Any = json.loads(open(_lowercase ).read() )
if not params:
raise ValueError(
f"""It seems that the json file at {parameter_file} is empty. Make sure you have a correct json file.""" )
if not args.output.endswith(""".pt""" ):
snake_case_ :Optional[int] = args.output + """.pt"""
snake_case_ :List[str] = OrderedDict()
with tf.device("""/CPU:0""" ):
snake_case_ :Dict = tf.train.load_checkpoint(args.tf_model_dir )
snake_case_ :str = reader.get_variable_to_shape_map()
for key_name in shapes.keys():
snake_case_ :List[Any] = reader.get_tensor(_lowercase ).astype(np.floataa )
if key_name.endswith("""/adam_m""" ) or key_name.endswith("""/adam_v""" ):
continue
if key_name.startswith("""pasts/""" ):
if key_name.startswith("""pasts/mlp""" ):
snake_case_ :Any = int(key_name[9] )
elif key_name.startswith("""pasts/out""" ):
snake_case_ :Optional[int] = 8
snake_case_ :List[str] = """model.sqout.%d.weight""" % (player * 2) # enter to nn.Sequencial with Tanh, so 2 at a time
snake_case_ :Optional[Any] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :List[str] = torch.tensor(_lowercase )
elif key_name.startswith("""model/moe""" ):
snake_case_ :Tuple = int(key_name[9:].split("""/""" )[0] )
if key_name.endswith("""/switch_gating/kernel""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.feed_forward.mlp.router.classifier.weight""" % player
snake_case_ :Optional[Any] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/softmlp/kernel""" ):
snake_case_ :List[Any] = """model.blocks.%d.feed_forward.soft_bypass_mlp.weight""" % player
snake_case_ :Optional[int] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/wo/kernel""" ) or key_name.endswith("""/wi/kernel""" ):
snake_case_ :Dict = key_name[-9:-7]
for i in range(16 ):
snake_case_ :str = """model.blocks.%d.feed_forward.mlp.experts.expert_%d.%s.weight""" % (player, i, nlayer)
snake_case_ :Tuple = (
vnp[i].transpose([1, 0] ).copy()
) # In Mesh-Tensorflow, it is one array, so it is divided
snake_case_ :Optional[int] = torch.tensor(_lowercase )
elif key_name.startswith("""model/mlp""" ):
snake_case_ :Optional[int] = int(key_name[9:].split("""/""" )[0] )
if key_name.endswith("""/p1/kernel""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.feed_forward.mlp.wi.weight""" % player
snake_case_ :Dict = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/p1/bias""" ):
snake_case_ :List[Any] = """model.blocks.%d.feed_forward.mlp.wi.bias""" % player
snake_case_ :str = vnp.copy() # same because it is one dimensional
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/p2/kernel""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.feed_forward.mlp.wo.weight""" % player
snake_case_ :Dict = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Tuple = torch.tensor(_lowercase )
elif key_name.endswith("""/p2/bias""" ):
snake_case_ :Dict = """model.blocks.%d.feed_forward.mlp.wo.bias""" % player
snake_case_ :Any = vnp.copy() # same because it is one dimensional
snake_case_ :Optional[int] = torch.tensor(_lowercase )
elif key_name.startswith("""model/ln""" ):
snake_case_ :Union[str, Any] = int(key_name[8:].split("""/""" )[0] )
if key_name.endswith("""/b""" ):
snake_case_ :str = """model.blocks.%d.feed_forward.norm.bias""" % player
snake_case_ :Dict = vnp.copy() # same because it is one dimensional
snake_case_ :int = torch.tensor(_lowercase )
elif key_name.endswith("""/g""" ):
snake_case_ :Dict = """model.blocks.%d.feed_forward.norm.weight""" % player
snake_case_ :Dict = vnp.copy() # same because it is one dimensional
snake_case_ :Tuple = torch.tensor(_lowercase )
elif key_name.startswith("""model/att""" ):
snake_case_ :List[str] = int(key_name[9:].split("""/""" )[0] )
if key_name.endswith("""/qkv/kernel""" ):
snake_case_ :Optional[int] = vnp.copy() # Compute same dimension as Mesh-tensorflow using einsum
snake_case_ :Dict = state[:, 0, :, :]
snake_case_ :int = state[:, 1, :, :]
snake_case_ :List[str] = state[:, 2, :, :]
snake_case_ :str = (
state_q.reshape([state_q.shape[0], state_q.shape[1] * state_q.shape[2]] )
.transpose([1, 0] )
.copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Any = (
state_k.reshape([state_k.shape[0], state_k.shape[1] * state_k.shape[2]] )
.transpose([1, 0] )
.copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[int] = (
state_v.reshape([state_v.shape[0], state_v.shape[1] * state_v.shape[2]] )
.transpose([1, 0] )
.copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :int = """model.blocks.%d.self_attn.self_attn.q_proj.weight""" % player
snake_case_ :int = torch.tensor(_lowercase )
snake_case_ :Optional[Any] = """model.blocks.%d.self_attn.self_attn.k_proj.weight""" % player
snake_case_ :Dict = torch.tensor(_lowercase )
snake_case_ :Dict = """model.blocks.%d.self_attn.self_attn.v_proj.weight""" % player
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/o/kernel""" ):
snake_case_ :str = """model.blocks.%d.self_attn.self_attn.out_proj.weight""" % player
snake_case_ :str = (
vnp.reshape([vnp.shape[0] * vnp.shape[1], vnp.shape[2]] ).transpose([1, 0] ).copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Any = torch.tensor(_lowercase )
elif key_name.startswith("""model/an""" ):
snake_case_ :Optional[int] = int(key_name[8:].split("""/""" )[0] )
if key_name.endswith("""/b""" ):
snake_case_ :Any = """model.blocks.%d.self_attn.norm.bias""" % player
snake_case_ :Optional[int] = vnp.copy() # same because it is one dimensional
snake_case_ :Tuple = torch.tensor(_lowercase )
elif key_name.endswith("""/g""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.self_attn.norm.weight""" % player
snake_case_ :Dict = vnp.copy() # same because it is one dimensional
snake_case_ :Optional[int] = torch.tensor(_lowercase )
elif (
key_name.startswith("""model/wte""" )
or key_name.startswith("""model/wpe""" )
or key_name.startswith("""model/ete""" )
):
snake_case_ :List[Any] = {"""wte""": """embed_tokens""", """wpe""": """position_embeddings""", """ete""": """extra_position_embeddings"""}[
key_name[-3:]
]
snake_case_ :Optional[Any] = """model.%s.weight""" % nlayer
snake_case_ :Any = vnp.copy() # same in embedded
snake_case_ :List[Any] = torch.tensor(_lowercase )
if key_name.startswith("""model/wte""" ):
snake_case_ :Tuple = """lm_head.weight"""
snake_case_ :List[str] = vnp.copy() # same in embedded
snake_case_ :List[Any] = torch.tensor(_lowercase )
elif key_name.startswith("""model/wob""" ):
snake_case_ :str = """final_logits_bias"""
snake_case_ :Any = vnp.copy() # same in embedded
snake_case_ :List[Any] = state.reshape((1, -1) )
snake_case_ :Union[str, Any] = torch.tensor(_lowercase )
elif key_name == "model/dense/kernel":
snake_case_ :str = """model.last_project.weight"""
snake_case_ :Dict = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :int = torch.tensor(_lowercase )
elif key_name == "model/dense_1/bias":
snake_case_ :Optional[int] = """model.last_project.bias"""
snake_case_ :Tuple = vnp.copy() # same because it is one dimensional
snake_case_ :Any = torch.tensor(_lowercase )
torch.save(_lowercase, args.output )
if __name__ == "__main__":
__a = argparse.ArgumentParser(
description="model converter.", formatter_class=argparse.ArgumentDefaultsHelpFormatter
)
parser.add_argument("--tf_model_dir", metavar="PATH", type=str, required=True, help="import model")
parser.add_argument("--output", metavar="PATH", type=str, required=True, help="output model")
__a = parser.parse_args()
convert_tf_gptsan_to_pt(args)
| 66 | 1 |
from __future__ import annotations
UpperCAmelCase__ = list[list[int]]
# assigning initial values to the grid
UpperCAmelCase__ = [
[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
UpperCAmelCase__ = [
[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 _a ( a :Matrix , a :int , a :int , a :int ) -> bool:
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 _a ( a :Matrix ) -> tuple[int, int] | None:
for i in range(9 ):
for j in range(9 ):
if grid[i][j] == 0:
return i, j
return None
def _a ( a :Matrix ) -> Matrix | None:
if location := find_empty_location(a ):
a , a = location
else:
# If the location is ``None``, then the grid is solved.
return grid
for digit in range(1 , 10 ):
if is_safe(a , a , a , a ):
a = digit
if sudoku(a ) is not None:
return grid
a = 0
return None
def _a ( a :Matrix ) -> None:
for row in grid:
for cell in row:
print(a , 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:")
UpperCAmelCase__ = sudoku(example_grid)
if solution is not None:
print_solution(solution)
else:
print("Cannot find a solution.")
| 0 |
"""simple docstring"""
import numpy as np
import pandas as pd
from sklearn.preprocessing import MinMaxScaler
from tensorflow.keras.layers import LSTM, Dense
from tensorflow.keras.models import Sequential
if __name__ == "__main__":
__a = pd.read_csv("sample_data.csv", header=None)
__a = df.shape[:1][0]
# If you're using some other dataset input the target column
__a = df.iloc[:, 1:2]
__a = actual_data.values.reshape(len_data, 1)
__a = MinMaxScaler().fit_transform(actual_data)
__a = 10
__a = 5
__a = 20
__a = len_data - periods * look_back
__a = actual_data[:division]
__a = actual_data[division - look_back :]
__a , __a = [], []
__a , __a = [], []
for i in range(0, len(train_data) - forward_days - look_back + 1):
train_x.append(train_data[i : i + look_back])
train_y.append(train_data[i + look_back : i + look_back + forward_days])
for i in range(0, len(test_data) - forward_days - look_back + 1):
test_x.append(test_data[i : i + look_back])
test_y.append(test_data[i + look_back : i + look_back + forward_days])
__a = np.array(train_x)
__a = np.array(test_x)
__a = np.array([list(i.ravel()) for i in train_y])
__a = np.array([list(i.ravel()) for i in test_y])
__a = Sequential()
model.add(LSTM(1_28, input_shape=(look_back, 1), return_sequences=True))
model.add(LSTM(64, input_shape=(1_28, 1)))
model.add(Dense(forward_days))
model.compile(loss="mean_squared_error", optimizer="adam")
__a = model.fit(
x_train, y_train, epochs=1_50, verbose=1, shuffle=True, batch_size=4
)
__a = model.predict(x_test)
| 66 | 0 |
'''simple docstring'''
from ...configuration_utils import PretrainedConfig
from ...utils import logging
from ...utils.backbone_utils import BackboneConfigMixin, get_aligned_output_features_output_indices
SCREAMING_SNAKE_CASE_: Optional[Any] =logging.get_logger(__name__)
SCREAMING_SNAKE_CASE_: List[Any] ={
'facebook/convnextv2-tiny-1k-224': 'https://huggingface.co/facebook/convnextv2-tiny-1k-224/resolve/main/config.json',
}
class __A ( UpperCamelCase__ , UpperCamelCase__ ):
a__ : int = """convnextv2"""
def __init__(self : Tuple , __a : Any=3 , __a : List[Any]=4 , __a : List[Any]=4 , __a : str=None , __a : Union[str, Any]=None , __a : Any="gelu" , __a : str=0.02 , __a : str=1E-12 , __a : str=0.0 , __a : str=224 , __a : Optional[Any]=None , __a : Union[str, Any]=None , **__a : Tuple , ):
super().__init__(**__a )
UpperCAmelCase_ = num_channels
UpperCAmelCase_ = patch_size
UpperCAmelCase_ = num_stages
UpperCAmelCase_ = [96, 192, 384, 768] if hidden_sizes is None else hidden_sizes
UpperCAmelCase_ = [3, 3, 9, 3] if depths is None else depths
UpperCAmelCase_ = hidden_act
UpperCAmelCase_ = initializer_range
UpperCAmelCase_ = layer_norm_eps
UpperCAmelCase_ = drop_path_rate
UpperCAmelCase_ = image_size
UpperCAmelCase_ = ["stem"] + [f"""stage{idx}""" for idx in range(1 , len(self.depths ) + 1 )]
UpperCAmelCase_ , UpperCAmelCase_ = get_aligned_output_features_output_indices(
out_features=__a , out_indices=__a , stage_names=self.stage_names )
| 1 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tokenizers_available, is_torch_available
__a = {
"configuration_altclip": [
"ALTCLIP_PRETRAINED_CONFIG_ARCHIVE_MAP",
"AltCLIPConfig",
"AltCLIPTextConfig",
"AltCLIPVisionConfig",
],
"processing_altclip": ["AltCLIPProcessor"],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"ALTCLIP_PRETRAINED_MODEL_ARCHIVE_LIST",
"AltCLIPPreTrainedModel",
"AltCLIPModel",
"AltCLIPTextModel",
"AltCLIPVisionModel",
]
if TYPE_CHECKING:
from .configuration_altclip import (
ALTCLIP_PRETRAINED_CONFIG_ARCHIVE_MAP,
AltCLIPConfig,
AltCLIPTextConfig,
AltCLIPVisionConfig,
)
from .processing_altclip import AltCLIPProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_altclip import (
ALTCLIP_PRETRAINED_MODEL_ARCHIVE_LIST,
AltCLIPModel,
AltCLIPPreTrainedModel,
AltCLIPTextModel,
AltCLIPVisionModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 | 0 |
'''simple docstring'''
from ...configuration_utils import PretrainedConfig
from ...utils import logging
lowerCamelCase : int = logging.get_logger(__name__)
lowerCamelCase : List[Any] = {
'google/realm-cc-news-pretrained-embedder': (
'https://huggingface.co/google/realm-cc-news-pretrained-embedder/resolve/main/config.json'
),
'google/realm-cc-news-pretrained-encoder': (
'https://huggingface.co/google/realm-cc-news-pretrained-encoder/resolve/main/config.json'
),
'google/realm-cc-news-pretrained-scorer': (
'https://huggingface.co/google/realm-cc-news-pretrained-scorer/resolve/main/config.json'
),
'google/realm-cc-news-pretrained-openqa': (
'https://huggingface.co/google/realm-cc-news-pretrained-openqa/aresolve/main/config.json'
),
'google/realm-orqa-nq-openqa': 'https://huggingface.co/google/realm-orqa-nq-openqa/resolve/main/config.json',
'google/realm-orqa-nq-reader': 'https://huggingface.co/google/realm-orqa-nq-reader/resolve/main/config.json',
'google/realm-orqa-wq-openqa': 'https://huggingface.co/google/realm-orqa-wq-openqa/resolve/main/config.json',
'google/realm-orqa-wq-reader': 'https://huggingface.co/google/realm-orqa-wq-reader/resolve/main/config.json',
# See all REALM models at https://huggingface.co/models?filter=realm
}
class __lowerCAmelCase (lowercase_ ):
'''simple docstring'''
lowerCAmelCase__ : Optional[int] = """realm"""
def __init__(self : str , UpperCamelCase : List[Any]=30522 , UpperCamelCase : List[Any]=768 , UpperCamelCase : int=128 , UpperCamelCase : Any=12 , UpperCamelCase : Tuple=12 , UpperCamelCase : List[Any]=8 , UpperCamelCase : Union[str, Any]=3072 , UpperCamelCase : List[str]="gelu_new" , UpperCamelCase : Any=0.1 , UpperCamelCase : List[str]=0.1 , UpperCamelCase : Dict=512 , UpperCamelCase : Dict=2 , UpperCamelCase : List[Any]=0.02 , UpperCamelCase : List[Any]=1E-12 , UpperCamelCase : Dict=256 , UpperCamelCase : Union[str, Any]=10 , UpperCamelCase : Optional[int]=1E-3 , UpperCamelCase : Tuple=5 , UpperCamelCase : Optional[int]=320 , UpperCamelCase : List[str]=13353718 , UpperCamelCase : Optional[Any]=5000 , UpperCamelCase : str=1 , UpperCamelCase : Union[str, Any]=0 , UpperCamelCase : List[Any]=2 , **UpperCamelCase : int , ):
'''simple docstring'''
super().__init__(pad_token_id=UpperCamelCase , bos_token_id=UpperCamelCase , eos_token_id=UpperCamelCase , **UpperCamelCase )
# Common config
lowercase__ = vocab_size
lowercase__ = max_position_embeddings
lowercase__ = hidden_size
lowercase__ = retriever_proj_size
lowercase__ = num_hidden_layers
lowercase__ = num_attention_heads
lowercase__ = num_candidates
lowercase__ = intermediate_size
lowercase__ = hidden_act
lowercase__ = hidden_dropout_prob
lowercase__ = attention_probs_dropout_prob
lowercase__ = initializer_range
lowercase__ = type_vocab_size
lowercase__ = layer_norm_eps
# Reader config
lowercase__ = span_hidden_size
lowercase__ = max_span_width
lowercase__ = reader_layer_norm_eps
lowercase__ = reader_beam_size
lowercase__ = reader_seq_len
# Retrieval config
lowercase__ = num_block_records
lowercase__ = searcher_beam_size
| 2 |
"""simple docstring"""
import argparse
import gdown
import numpy as np
import torch
from huggingface_hub import hf_hub_download
from transformers import (
CLIPTokenizer,
CLIPTokenizerFast,
VideoMAEImageProcessor,
XCLIPConfig,
XCLIPModel,
XCLIPProcessor,
XCLIPTextConfig,
XCLIPVisionConfig,
)
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :int = XCLIPTextConfig()
# derive patch size from model name
snake_case_ :Union[str, Any] = model_name.find("""patch""" )
snake_case_ :List[str] = int(model_name[start_idx + len("""patch""" ) : start_idx + len("""patch""" ) + 2] )
snake_case_ :Any = XCLIPVisionConfig(patch_size=_lowercase, num_frames=_lowercase )
if "large" in model_name:
snake_case_ :Optional[Any] = 768
snake_case_ :Union[str, Any] = 3072
snake_case_ :Any = 12
snake_case_ :Any = 1024
snake_case_ :str = 4096
snake_case_ :Union[str, Any] = 16
snake_case_ :Union[str, Any] = 24
snake_case_ :Tuple = 768
snake_case_ :Any = 3072
if model_name == "xclip-large-patch14-16-frames":
snake_case_ :Any = 336
snake_case_ :Any = XCLIPConfig.from_text_vision_configs(_lowercase, _lowercase )
if "large" in model_name:
snake_case_ :List[Any] = 768
return config
def A_ ( _lowercase ):
'''simple docstring'''
if name == "token_embedding.weight":
snake_case_ :Optional[Any] = name.replace("""token_embedding.weight""", """text_model.embeddings.token_embedding.weight""" )
if name == "positional_embedding":
snake_case_ :Tuple = name.replace("""positional_embedding""", """text_model.embeddings.position_embedding.weight""" )
if "ln_1" in name:
snake_case_ :Dict = name.replace("""ln_1""", """layer_norm1""" )
if "ln_2" in name:
snake_case_ :str = name.replace("""ln_2""", """layer_norm2""" )
if "c_fc" in name:
snake_case_ :str = name.replace("""c_fc""", """fc1""" )
if "c_proj" in name:
snake_case_ :int = name.replace("""c_proj""", """fc2""" )
if name.startswith("""transformer.resblocks""" ):
snake_case_ :Union[str, Any] = name.replace("""transformer.resblocks""", """text_model.encoder.layers""" )
if "attn.out_proj" in name and "message" not in name:
snake_case_ :Union[str, Any] = name.replace("""attn.out_proj""", """self_attn.out_proj""" )
if "ln_final" in name:
snake_case_ :Union[str, Any] = name.replace("""ln_final""", """text_model.final_layer_norm""" )
# visual encoder
if name == "visual.class_embedding":
snake_case_ :Any = name.replace("""visual.class_embedding""", """vision_model.embeddings.class_embedding""" )
if name == "visual.positional_embedding":
snake_case_ :Optional[int] = name.replace("""visual.positional_embedding""", """vision_model.embeddings.position_embedding.weight""" )
if name.startswith("""visual.transformer.resblocks""" ):
snake_case_ :Union[str, Any] = name.replace("""visual.transformer.resblocks""", """vision_model.encoder.layers""" )
if "visual.conv1" in name:
snake_case_ :int = name.replace("""visual.conv1""", """vision_model.embeddings.patch_embedding""" )
if "visual.ln_pre" in name:
snake_case_ :Any = name.replace("""visual.ln_pre""", """vision_model.pre_layernorm""" )
if "visual.ln_post" in name:
snake_case_ :str = name.replace("""visual.ln_post""", """vision_model.post_layernorm""" )
if "visual.proj" in name:
snake_case_ :Union[str, Any] = name.replace("""visual.proj""", """visual_projection.weight""" )
if "text_projection" in name:
snake_case_ :Dict = name.replace("""text_projection""", """text_projection.weight""" )
# things on top
if "prompts_visual_proj" in name:
snake_case_ :List[str] = name.replace("""prompts_visual_proj""", """prompts_visual_projection""" )
if "prompts_visual_ln" in name:
snake_case_ :Dict = name.replace("""prompts_visual_ln""", """prompts_visual_layernorm""" )
# mit
if name == "mit.positional_embedding":
snake_case_ :str = name.replace("""positional""", """position""" )
if name.startswith("""mit.resblocks""" ):
snake_case_ :Dict = name.replace("""mit.resblocks""", """mit.encoder.layers""" )
# prompts generator
if name.startswith("""prompts_generator.norm""" ):
snake_case_ :Union[str, Any] = name.replace("""prompts_generator.norm""", """prompts_generator.layernorm""" )
return name
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
for key in orig_state_dict.copy().keys():
snake_case_ :Dict = orig_state_dict.pop(_lowercase )
if "attn.in_proj" in key:
snake_case_ :Optional[Any] = key.split(""".""" )
if key.startswith("""visual""" ):
snake_case_ :Any = key_split[3]
snake_case_ :Optional[Any] = config.vision_config.hidden_size
if "message_attn" in key:
if "weight" in key:
snake_case_ :str = val[
:dim, :
]
snake_case_ :Optional[int] = val[
dim : dim * 2, :
]
snake_case_ :Union[str, Any] = val[
-dim:, :
]
else:
snake_case_ :Dict = val[
:dim
]
snake_case_ :Optional[int] = val[
dim : dim * 2
]
snake_case_ :Optional[int] = val[
-dim:
]
else:
if "weight" in key:
snake_case_ :Optional[Any] = val[
:dim, :
]
snake_case_ :List[str] = val[
dim : dim * 2, :
]
snake_case_ :Dict = val[
-dim:, :
]
else:
snake_case_ :Union[str, Any] = val[:dim]
snake_case_ :Union[str, Any] = val[
dim : dim * 2
]
snake_case_ :Union[str, Any] = val[-dim:]
elif key.startswith("""mit""" ):
snake_case_ :Tuple = key_split[2]
snake_case_ :Union[str, Any] = config.vision_config.mit_hidden_size
if "weight" in key:
snake_case_ :Optional[int] = val[:dim, :]
snake_case_ :Optional[int] = val[dim : dim * 2, :]
snake_case_ :str = val[-dim:, :]
else:
snake_case_ :str = val[:dim]
snake_case_ :Any = val[dim : dim * 2]
snake_case_ :int = val[-dim:]
else:
snake_case_ :Tuple = key_split[2]
snake_case_ :Any = config.text_config.hidden_size
if "weight" in key:
snake_case_ :Dict = val[:dim, :]
snake_case_ :Dict = val[
dim : dim * 2, :
]
snake_case_ :List[str] = val[-dim:, :]
else:
snake_case_ :Any = val[:dim]
snake_case_ :Tuple = val[
dim : dim * 2
]
snake_case_ :List[str] = val[-dim:]
else:
snake_case_ :Optional[int] = rename_key(_lowercase )
if new_key_name in ["visual_projection.weight", "text_projection.weight"]:
snake_case_ :Optional[Any] = val.T
snake_case_ :Tuple = val
return orig_state_dict
def A_ ( _lowercase ):
'''simple docstring'''
if num_frames == 8:
snake_case_ :str = """eating_spaghetti_8_frames.npy"""
elif num_frames == 16:
snake_case_ :int = """eating_spaghetti.npy"""
elif num_frames == 32:
snake_case_ :List[str] = """eating_spaghetti_32_frames.npy"""
snake_case_ :int = hf_hub_download(
repo_id="""hf-internal-testing/spaghetti-video""", filename=_lowercase, repo_type="""dataset""", )
snake_case_ :Union[str, Any] = np.load(_lowercase )
return list(_lowercase )
def A_ ( _lowercase, _lowercase=None, _lowercase=False ):
'''simple docstring'''
snake_case_ :List[Any] = {
# fully supervised kinetics-400 checkpoints
"""xclip-base-patch32""": """https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_32_8.pth""",
"""xclip-base-patch32-16-frames""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_32_16.pth"""
),
"""xclip-base-patch16""": """https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_16_8.pth""",
"""xclip-base-patch16-16-frames""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_16_16.pth"""
),
"""xclip-large-patch14""": """https://drive.google.com/u/0/uc?id=1NUOImq0o5DlQTST17iIP3vG7DgmHQuCx&export=download&confirm=t&uuid=b26caedc-88e2-473e-830a-9d158b653cdb""",
"""xclip-large-patch14-16-frames""": """https://drive.google.com/u/0/uc?id=1FOYgnJc097OJ4lGwtRCCydQyVPJEOH7d&export=download&confirm=t&uuid=538fa810-e671-4050-b385-9a623f89804f""",
# fully supervised kinetics-600 checkpoints
"""xclip-base-patch16-kinetics-600""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k600_16_8.pth"""
),
"""xclip-base-patch16-kinetics-600-16-frames""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k600_16_16.pth"""
),
"""xclip-large-patch14-kinetics-600""": """https://drive.google.com/u/0/uc?id=1FV8C1INuM91sLAN4ImjzePLIlpMSihwV&export=download&confirm=t&uuid=141d4977-4a65-44ae-864f-4b0c19f838be""",
# few shot
"""xclip-base-patch16-hmdb-2-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_2.pth"""
),
"""xclip-base-patch16-hmdb-4-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_4.pth"""
),
"""xclip-base-patch16-hmdb-8-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_8.pth"""
),
"""xclip-base-patch16-hmdb-16-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_16.pth"""
),
"""xclip-base-patch16-ucf-2-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_2.pth"""
),
"""xclip-base-patch16-ucf-4-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_4.pth"""
),
"""xclip-base-patch16-ucf-8-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_8.pth"""
),
"""xclip-base-patch16-ucf-16-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_16.pth"""
),
# zero shot
"""xclip-base-patch16-zero-shot""": """https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/zero.pth""",
}
snake_case_ :Optional[int] = model_to_url[model_name]
snake_case_ :int = 8
if "16-frames" in model_name:
snake_case_ :List[Any] = 16
elif "shot" in model_name:
snake_case_ :Dict = 32
snake_case_ :Optional[int] = get_xclip_config(_lowercase, _lowercase )
snake_case_ :Optional[Any] = XCLIPModel(_lowercase )
model.eval()
if "drive" in checkpoint_url:
snake_case_ :List[str] = """pytorch_model.bin"""
gdown.cached_download(_lowercase, _lowercase, quiet=_lowercase )
snake_case_ :List[Any] = torch.load(_lowercase, map_location="""cpu""" )["""model"""]
else:
snake_case_ :Tuple = torch.hub.load_state_dict_from_url(_lowercase )["""model"""]
snake_case_ :Union[str, Any] = convert_state_dict(_lowercase, _lowercase )
snake_case_ :str = XCLIPModel(_lowercase )
snake_case_, snake_case_ :Optional[int] = model.load_state_dict(_lowercase, strict=_lowercase )
assert missing_keys == ["text_model.embeddings.position_ids", "vision_model.embeddings.position_ids"]
model.eval()
snake_case_ :List[str] = 336 if model_name == """xclip-large-patch14-16-frames""" else 224
snake_case_ :List[Any] = VideoMAEImageProcessor(size=_lowercase )
snake_case_ :Any = CLIPTokenizer.from_pretrained("""openai/clip-vit-base-patch32""" )
snake_case_ :str = CLIPTokenizerFast.from_pretrained("""openai/clip-vit-base-patch32""" )
snake_case_ :Optional[Any] = XCLIPProcessor(image_processor=_lowercase, tokenizer=_lowercase )
snake_case_ :Optional[int] = prepare_video(_lowercase )
snake_case_ :Optional[Any] = processor(
text=["""playing sports""", """eating spaghetti""", """go shopping"""], videos=_lowercase, return_tensors="""pt""", padding=_lowercase )
print("""Shape of pixel values:""", inputs.pixel_values.shape )
with torch.no_grad():
snake_case_ :List[Any] = model(**_lowercase )
# Verify outputs
snake_case_ :List[Any] = outputs.logits_per_video
snake_case_ :Any = logits_per_video.softmax(dim=1 )
print("""Probs:""", _lowercase )
# kinetics-400
if model_name == "xclip-base-patch32":
snake_case_ :Union[str, Any] = torch.tensor([[0.0019, 0.9951, 0.0030]] )
elif model_name == "xclip-base-patch32-16-frames":
snake_case_ :str = torch.tensor([[7.09_99e-04, 9.98_83e-01, 4.55_80e-04]] )
elif model_name == "xclip-base-patch16":
snake_case_ :Tuple = torch.tensor([[0.0083, 0.9681, 0.0236]] )
elif model_name == "xclip-base-patch16-16-frames":
snake_case_ :Any = torch.tensor([[7.69_37e-04, 9.97_28e-01, 1.94_73e-03]] )
elif model_name == "xclip-large-patch14":
snake_case_ :str = torch.tensor([[0.0062, 0.9864, 0.0075]] )
elif model_name == "xclip-large-patch14-16-frames":
snake_case_ :Tuple = torch.tensor([[3.38_77e-04, 9.99_37e-01, 2.88_88e-04]] )
# kinetics-600
elif model_name == "xclip-base-patch16-kinetics-600":
snake_case_ :List[Any] = torch.tensor([[0.0555, 0.8914, 0.0531]] )
elif model_name == "xclip-base-patch16-kinetics-600-16-frames":
snake_case_ :Union[str, Any] = torch.tensor([[3.85_54e-04, 9.99_29e-01, 3.27_54e-04]] )
elif model_name == "xclip-large-patch14-kinetics-600":
snake_case_ :List[Any] = torch.tensor([[0.0036, 0.9920, 0.0045]] )
# few shot
elif model_name == "xclip-base-patch16-hmdb-2-shot":
snake_case_ :Dict = torch.tensor([[7.18_90e-06, 9.99_94e-01, 5.65_59e-05]] )
elif model_name == "xclip-base-patch16-hmdb-4-shot":
snake_case_ :Union[str, Any] = torch.tensor([[1.03_20e-05, 9.99_93e-01, 6.24_35e-05]] )
elif model_name == "xclip-base-patch16-hmdb-8-shot":
snake_case_ :str = torch.tensor([[4.13_77e-06, 9.99_90e-01, 9.83_86e-05]] )
elif model_name == "xclip-base-patch16-hmdb-16-shot":
snake_case_ :str = torch.tensor([[4.13_47e-05, 9.99_62e-01, 3.34_11e-04]] )
elif model_name == "xclip-base-patch16-ucf-2-shot":
snake_case_ :int = torch.tensor([[8.58_57e-05, 9.99_28e-01, 6.32_91e-04]] )
elif model_name == "xclip-base-patch16-ucf-4-shot":
snake_case_ :Optional[int] = torch.tensor([[8.58_57e-05, 9.99_28e-01, 6.32_91e-04]] )
elif model_name == "xclip-base-patch16-ucf-8-shot":
snake_case_ :Any = torch.tensor([[0.0027, 0.9904, 0.0070]] )
elif model_name == "xclip-base-patch16-ucf-16-shot":
snake_case_ :Tuple = torch.tensor([[9.82_19e-04, 9.95_93e-01, 3.08_63e-03]] )
# zero shot
elif model_name == "xclip-base-patch16-zero-shot":
snake_case_ :Union[str, Any] = torch.tensor([[3.50_82e-04, 9.97_85e-01, 1.79_66e-03]] )
else:
raise ValueError(f"""Model name {model_name} not supported""" )
assert torch.allclose(_lowercase, _lowercase, atol=1e-3 )
print("""Looks ok!""" )
if pytorch_dump_folder_path is not None:
print(f"""Saving model {model_name} to {pytorch_dump_folder_path}""" )
model.save_pretrained(_lowercase )
if push_to_hub:
print("""Pushing model, processor and slow tokenizer files to the hub...""" )
model.push_to_hub(_lowercase, organization="""nielsr""" )
processor.push_to_hub(_lowercase, organization="""nielsr""" )
slow_tokenizer.push_to_hub(_lowercase, organization="""nielsr""" )
if __name__ == "__main__":
__a = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--model_name",
default="xclip-base-patch32",
type=str,
help="Name of the model.",
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model directory."
)
parser.add_argument(
"--push_to_hub", action="store_true", help="Whether or not to push the converted model to the 🤗 hub."
)
__a = parser.parse_args()
convert_xclip_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 66 | 0 |
'''simple docstring'''
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available
lowercase : Any = {
'configuration_bigbird_pegasus': [
'BIGBIRD_PEGASUS_PRETRAINED_CONFIG_ARCHIVE_MAP',
'BigBirdPegasusConfig',
'BigBirdPegasusOnnxConfig',
],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
lowercase : Optional[Any] = [
'BIGBIRD_PEGASUS_PRETRAINED_MODEL_ARCHIVE_LIST',
'BigBirdPegasusForCausalLM',
'BigBirdPegasusForConditionalGeneration',
'BigBirdPegasusForQuestionAnswering',
'BigBirdPegasusForSequenceClassification',
'BigBirdPegasusModel',
'BigBirdPegasusPreTrainedModel',
]
if TYPE_CHECKING:
from .configuration_bigbird_pegasus import (
BIGBIRD_PEGASUS_PRETRAINED_CONFIG_ARCHIVE_MAP,
BigBirdPegasusConfig,
BigBirdPegasusOnnxConfig,
)
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_bigbird_pegasus import (
BIGBIRD_PEGASUS_PRETRAINED_MODEL_ARCHIVE_LIST,
BigBirdPegasusForCausalLM,
BigBirdPegasusForConditionalGeneration,
BigBirdPegasusForQuestionAnswering,
BigBirdPegasusForSequenceClassification,
BigBirdPegasusModel,
BigBirdPegasusPreTrainedModel,
)
else:
import sys
lowercase : Any = _LazyModule(__name__, globals()['__file__'], _import_structure, module_spec=__spec__)
| 3 |
"""simple docstring"""
import unittest
import numpy as np
from transformers import BertConfig, is_flax_available
from transformers.testing_utils import require_flax, slow
from ...test_modeling_flax_common import FlaxModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask
if is_flax_available():
from transformers.models.bert.modeling_flax_bert import (
FlaxBertForMaskedLM,
FlaxBertForMultipleChoice,
FlaxBertForNextSentencePrediction,
FlaxBertForPreTraining,
FlaxBertForQuestionAnswering,
FlaxBertForSequenceClassification,
FlaxBertForTokenClassification,
FlaxBertModel,
)
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def __init__( self: List[Any] , snake_case: List[str] , snake_case: Optional[Any]=13 , snake_case: List[str]=7 , snake_case: Dict=True , snake_case: List[str]=True , snake_case: Optional[int]=True , snake_case: Any=True , snake_case: Optional[Any]=99 , snake_case: Tuple=32 , snake_case: Tuple=5 , snake_case: Dict=4 , snake_case: Optional[Any]=37 , snake_case: Union[str, Any]="gelu" , snake_case: Tuple=0.1 , snake_case: List[Any]=0.1 , snake_case: List[str]=512 , snake_case: Optional[int]=16 , snake_case: int=2 , snake_case: List[Any]=0.0_2 , snake_case: Union[str, Any]=4 , ) -> List[str]:
snake_case_ :Dict = parent
snake_case_ :Any = batch_size
snake_case_ :Any = seq_length
snake_case_ :List[str] = is_training
snake_case_ :Optional[Any] = use_attention_mask
snake_case_ :Dict = use_token_type_ids
snake_case_ :Union[str, Any] = use_labels
snake_case_ :str = vocab_size
snake_case_ :int = hidden_size
snake_case_ :List[str] = num_hidden_layers
snake_case_ :Dict = num_attention_heads
snake_case_ :Any = intermediate_size
snake_case_ :Tuple = hidden_act
snake_case_ :int = hidden_dropout_prob
snake_case_ :Optional[Any] = attention_probs_dropout_prob
snake_case_ :Any = max_position_embeddings
snake_case_ :Union[str, Any] = type_vocab_size
snake_case_ :Optional[int] = type_sequence_label_size
snake_case_ :Union[str, Any] = initializer_range
snake_case_ :Tuple = num_choices
def lowerCAmelCase_ ( self: Tuple ) -> str:
snake_case_ :Optional[Any] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case_ :Union[str, Any] = None
if self.use_attention_mask:
snake_case_ :str = random_attention_mask([self.batch_size, self.seq_length] )
snake_case_ :Any = None
if self.use_token_type_ids:
snake_case_ :List[str] = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
snake_case_ :int = BertConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , is_decoder=snake_case , initializer_range=self.initializer_range , )
return config, input_ids, token_type_ids, attention_mask
def lowerCAmelCase_ ( self: Optional[int] ) -> int:
snake_case_ :str = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_, snake_case_ :Optional[int] = config_and_inputs
snake_case_ :Union[str, Any] = {"""input_ids""": input_ids, """token_type_ids""": token_type_ids, """attention_mask""": attention_mask}
return config, inputs_dict
def lowerCAmelCase_ ( self: Optional[Any] ) -> Any:
snake_case_ :int = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_, snake_case_ :Dict = config_and_inputs
snake_case_ :Union[str, Any] = True
snake_case_ :Optional[int] = floats_tensor([self.batch_size, self.seq_length, self.hidden_size] )
snake_case_ :Tuple = ids_tensor([self.batch_size, self.seq_length] , vocab_size=2 )
return (
config,
input_ids,
attention_mask,
encoder_hidden_states,
encoder_attention_mask,
)
@require_flax
class lowerCamelCase ( _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[str] = True
_A : Dict = (
(
FlaxBertModel,
FlaxBertForPreTraining,
FlaxBertForMaskedLM,
FlaxBertForMultipleChoice,
FlaxBertForQuestionAnswering,
FlaxBertForNextSentencePrediction,
FlaxBertForSequenceClassification,
FlaxBertForTokenClassification,
FlaxBertForQuestionAnswering,
)
if is_flax_available()
else ()
)
def lowerCAmelCase_ ( self: int ) -> List[str]:
snake_case_ :Any = FlaxBertModelTester(self )
@slow
def lowerCAmelCase_ ( self: List[str] ) -> Dict:
# Only check this for base model, not necessary for all model classes.
# This will also help speed-up tests.
snake_case_ :Dict = FlaxBertModel.from_pretrained("""bert-base-cased""" )
snake_case_ :Dict = model(np.ones((1, 1) ) )
self.assertIsNotNone(snake_case )
| 66 | 0 |
'''simple docstring'''
import unittest
from datasets import load_dataset
from transformers.pipelines import pipeline
from transformers.testing_utils import is_pipeline_test, nested_simplify, require_torch, slow
@is_pipeline_test
@require_torch
class UpperCAmelCase_ ( unittest.TestCase ):
@require_torch
def __UpperCAmelCase ( self : Any ) -> Optional[Any]:
lowerCAmelCase = pipeline(
task='zero-shot-audio-classification' , model='hf-internal-testing/tiny-clap-htsat-unfused' )
lowerCAmelCase = load_dataset('ashraq/esc50' )
lowerCAmelCase = dataset['train']['audio'][-1]['array']
lowerCAmelCase = audio_classifier(UpperCAmelCase__ , candidate_labels=['Sound of a dog', 'Sound of vaccum cleaner'] )
self.assertEqual(
nested_simplify(UpperCAmelCase__ ) , [{'score': 0.501, 'label': 'Sound of a dog'}, {'score': 0.499, 'label': 'Sound of vaccum cleaner'}] , )
@unittest.skip('No models are available in TF' )
def __UpperCAmelCase ( self : Any ) -> Optional[Any]:
pass
@slow
@require_torch
def __UpperCAmelCase ( self : Tuple ) -> List[Any]:
lowerCAmelCase = pipeline(
task='zero-shot-audio-classification' , model='laion/clap-htsat-unfused' , )
# This is an audio of a dog
lowerCAmelCase = load_dataset('ashraq/esc50' )
lowerCAmelCase = dataset['train']['audio'][-1]['array']
lowerCAmelCase = audio_classifier(UpperCAmelCase__ , candidate_labels=['Sound of a dog', 'Sound of vaccum cleaner'] )
self.assertEqual(
nested_simplify(UpperCAmelCase__ ) , [
{'score': 0.999, 'label': 'Sound of a dog'},
{'score': 0.001, 'label': 'Sound of vaccum cleaner'},
] , )
lowerCAmelCase = audio_classifier([audio] * 5 , candidate_labels=['Sound of a dog', 'Sound of vaccum cleaner'] )
self.assertEqual(
nested_simplify(UpperCAmelCase__ ) , [
[
{'score': 0.999, 'label': 'Sound of a dog'},
{'score': 0.001, 'label': 'Sound of vaccum cleaner'},
],
]
* 5 , )
lowerCAmelCase = audio_classifier(
[audio] * 5 , candidate_labels=['Sound of a dog', 'Sound of vaccum cleaner'] , batch_size=5 )
self.assertEqual(
nested_simplify(UpperCAmelCase__ ) , [
[
{'score': 0.999, 'label': 'Sound of a dog'},
{'score': 0.001, 'label': 'Sound of vaccum cleaner'},
],
]
* 5 , )
@unittest.skip('No models are available in TF' )
def __UpperCAmelCase ( self : Any ) -> Union[str, Any]:
pass
| 4 |
"""simple docstring"""
import math
class lowerCamelCase :
'''simple docstring'''
def lowerCAmelCase_ ( self: Tuple , snake_case: list[list[float]] , snake_case: list[int] ) -> int:
snake_case_ :Any = 0.0
snake_case_ :Tuple = 0.0
for i in range(len(snake_case ) ):
da += math.pow((sample[i] - weights[0][i]) , 2 )
da += math.pow((sample[i] - weights[1][i]) , 2 )
return 0 if da > da else 1
return 0
def lowerCAmelCase_ ( self: Optional[int] , snake_case: list[list[int | float]] , snake_case: list[int] , snake_case: int , snake_case: float ) -> list[list[int | float]]:
for i in range(len(snake_case ) ):
weights[j][i] += alpha * (sample[i] - weights[j][i])
return weights
def A_ ( ):
'''simple docstring'''
snake_case_ :Dict = [[1, 1, 0, 0], [0, 0, 0, 1], [1, 0, 0, 0], [0, 0, 1, 1]]
# weight initialization ( n, C )
snake_case_ :List[Any] = [[0.2, 0.6, 0.5, 0.9], [0.8, 0.4, 0.7, 0.3]]
# training
snake_case_ :Optional[Any] = SelfOrganizingMap()
snake_case_ :Dict = 3
snake_case_ :Dict = 0.5
for _ in range(_lowercase ):
for j in range(len(_lowercase ) ):
# training sample
snake_case_ :List[Any] = training_samples[j]
# Compute the winning vector
snake_case_ :Optional[int] = self_organizing_map.get_winner(_lowercase, _lowercase )
# Update the winning vector
snake_case_ :List[str] = self_organizing_map.update(_lowercase, _lowercase, _lowercase, _lowercase )
# classify test sample
snake_case_ :str = [0, 0, 0, 1]
snake_case_ :List[Any] = self_organizing_map.get_winner(_lowercase, _lowercase )
# results
print(f"""Clusters that the test sample belongs to : {winner}""" )
print(f"""Weights that have been trained : {weights}""" )
# running the main() function
if __name__ == "__main__":
main()
| 66 | 0 |
import warnings
from ...utils import logging
from .image_processing_imagegpt import ImageGPTImageProcessor
UpperCAmelCase__ = logging.get_logger(__name__)
class lowerCamelCase__ ( lowerCAmelCase):
def __init__(self , *UpperCAmelCase , **UpperCAmelCase ) -> None:
warnings.warn(
'''The class ImageGPTFeatureExtractor is deprecated and will be removed in version 5 of Transformers.'''
''' Please use ImageGPTImageProcessor instead.''' , UpperCAmelCase , )
super().__init__(*UpperCAmelCase , **UpperCAmelCase )
| 5 |
"""simple docstring"""
import collections
import inspect
import unittest
from transformers import SwinvaConfig
from transformers.testing_utils import require_torch, require_vision, slow, torch_device
from transformers.utils import cached_property, is_torch_available, is_vision_available
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, _config_zero_init, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from torch import nn
from transformers import SwinvaForImageClassification, SwinvaForMaskedImageModeling, SwinvaModel
from transformers.models.swinva.modeling_swinva import SWINV2_PRETRAINED_MODEL_ARCHIVE_LIST
if is_vision_available():
from PIL import Image
from transformers import AutoImageProcessor
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Optional[int] , snake_case: Any , snake_case: Optional[Any]=13 , snake_case: Tuple=32 , snake_case: Optional[int]=2 , snake_case: Tuple=3 , snake_case: Tuple=16 , snake_case: Optional[Any]=[1, 2, 1] , snake_case: Optional[int]=[2, 2, 4] , snake_case: Optional[int]=2 , snake_case: int=2.0 , snake_case: Union[str, Any]=True , snake_case: List[str]=0.0 , snake_case: List[Any]=0.0 , snake_case: Optional[Any]=0.1 , snake_case: List[Any]="gelu" , snake_case: Optional[int]=False , snake_case: Union[str, Any]=True , snake_case: Union[str, Any]=0.0_2 , snake_case: Optional[int]=1E-5 , snake_case: Optional[Any]=True , snake_case: List[Any]=None , snake_case: List[Any]=True , snake_case: Optional[Any]=10 , snake_case: str=8 , ) -> Tuple:
snake_case_ :Dict = parent
snake_case_ :Any = batch_size
snake_case_ :List[Any] = image_size
snake_case_ :List[Any] = patch_size
snake_case_ :int = num_channels
snake_case_ :Tuple = embed_dim
snake_case_ :str = depths
snake_case_ :str = num_heads
snake_case_ :Optional[int] = window_size
snake_case_ :Tuple = mlp_ratio
snake_case_ :Any = qkv_bias
snake_case_ :List[Any] = hidden_dropout_prob
snake_case_ :Optional[Any] = attention_probs_dropout_prob
snake_case_ :Union[str, Any] = drop_path_rate
snake_case_ :Any = hidden_act
snake_case_ :Optional[Any] = use_absolute_embeddings
snake_case_ :Union[str, Any] = patch_norm
snake_case_ :Dict = layer_norm_eps
snake_case_ :str = initializer_range
snake_case_ :Tuple = is_training
snake_case_ :Tuple = scope
snake_case_ :Union[str, Any] = use_labels
snake_case_ :Optional[Any] = type_sequence_label_size
snake_case_ :Dict = encoder_stride
def lowerCAmelCase_ ( self: int ) -> int:
snake_case_ :List[str] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case_ :Any = None
if self.use_labels:
snake_case_ :str = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ :int = self.get_config()
return config, pixel_values, labels
def lowerCAmelCase_ ( self: str ) -> Union[str, Any]:
return SwinvaConfig(
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 , )
def lowerCAmelCase_ ( self: str , snake_case: Optional[int] , snake_case: Dict , snake_case: str ) -> List[Any]:
snake_case_ :Union[str, Any] = SwinvaModel(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Optional[int] = model(snake_case )
snake_case_ :Optional[Any] = ((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 lowerCAmelCase_ ( self: int , snake_case: List[str] , snake_case: Tuple , snake_case: int ) -> Any:
snake_case_ :Dict = SwinvaForMaskedImageModeling(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Tuple = model(snake_case )
self.parent.assertEqual(
result.logits.shape , (self.batch_size, self.num_channels, self.image_size, self.image_size) )
# test greyscale images
snake_case_ :List[Any] = 1
snake_case_ :int = SwinvaForMaskedImageModeling(snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Union[str, Any] = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] )
snake_case_ :int = model(snake_case )
self.parent.assertEqual(result.logits.shape , (self.batch_size, 1, self.image_size, self.image_size) )
def lowerCAmelCase_ ( self: List[Any] , snake_case: Any , snake_case: List[str] , snake_case: Union[str, Any] ) -> Tuple:
snake_case_ :int = self.type_sequence_label_size
snake_case_ :List[Any] = SwinvaForImageClassification(snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Dict = model(snake_case , labels=snake_case )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
def lowerCAmelCase_ ( self: int ) -> str:
snake_case_ :Any = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_ :List[str] = config_and_inputs
snake_case_ :List[Any] = {"""pixel_values""": pixel_values}
return config, inputs_dict
@require_torch
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : Optional[Any] = (
(SwinvaModel, SwinvaForImageClassification, SwinvaForMaskedImageModeling) if is_torch_available() else ()
)
_A : Any = (
{"""feature-extraction""": SwinvaModel, """image-classification""": SwinvaForImageClassification}
if is_torch_available()
else {}
)
_A : List[Any] = False
_A : List[str] = False
_A : Tuple = False
_A : List[str] = False
def lowerCAmelCase_ ( self: Dict ) -> List[Any]:
snake_case_ :Optional[int] = SwinvaModelTester(self )
snake_case_ :List[str] = ConfigTester(self , config_class=snake_case , embed_dim=37 )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> List[Any]:
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 lowerCAmelCase_ ( self: Union[str, Any] ) -> Tuple:
snake_case_ :List[str] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case )
@unittest.skip(reason="""Got `CUDA error: misaligned address` with PyTorch 2.0.0.""" )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> str:
pass
@unittest.skip(reason="""Swinv2 does not use inputs_embeds""" )
def lowerCAmelCase_ ( self: int ) -> Dict:
pass
def lowerCAmelCase_ ( self: List[str] ) -> Union[str, Any]:
snake_case_, snake_case_ :List[str] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :Optional[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 lowerCAmelCase_ ( self: Dict ) -> Optional[int]:
snake_case_, snake_case_ :Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :Optional[int] = model_class(snake_case )
snake_case_ :List[Any] = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case_ :int = [*signature.parameters.keys()]
snake_case_ :List[Any] = ["""pixel_values"""]
self.assertListEqual(arg_names[:1] , snake_case )
def lowerCAmelCase_ ( self: List[str] ) -> Optional[Any]:
snake_case_, snake_case_ :List[str] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :List[str] = True
for model_class in self.all_model_classes:
snake_case_ :List[Any] = True
snake_case_ :Any = False
snake_case_ :Optional[int] = True
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_ :str = outputs.attentions
snake_case_ :Dict = len(self.model_tester.depths )
self.assertEqual(len(snake_case ) , snake_case )
# check that output_attentions also work using config
del inputs_dict["output_attentions"]
snake_case_ :Union[str, Any] = True
snake_case_ :Tuple = config.window_size**2
snake_case_ :Any = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :Union[str, Any] = model(**self._prepare_for_class(snake_case , snake_case ) )
snake_case_ :int = outputs.attentions
self.assertEqual(len(snake_case ) , snake_case )
self.assertListEqual(
list(attentions[0].shape[-3:] ) , [self.model_tester.num_heads[0], window_size_squared, window_size_squared] , )
snake_case_ :Any = len(snake_case )
# Check attention is always last and order is fine
snake_case_ :int = True
snake_case_ :Dict = True
snake_case_ :Optional[int] = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :Dict = model(**self._prepare_for_class(snake_case , snake_case ) )
if hasattr(self.model_tester , """num_hidden_states_types""" ):
snake_case_ :Any = self.model_tester.num_hidden_states_types
else:
# also another +1 for reshaped_hidden_states
snake_case_ :int = 2
self.assertEqual(out_len + added_hidden_states , len(snake_case ) )
snake_case_ :str = outputs.attentions
self.assertEqual(len(snake_case ) , snake_case )
self.assertListEqual(
list(self_attentions[0].shape[-3:] ) , [self.model_tester.num_heads[0], window_size_squared, window_size_squared] , )
def lowerCAmelCase_ ( self: int , snake_case: Dict , snake_case: Dict , snake_case: Optional[Any] , snake_case: Dict ) -> List[str]:
snake_case_ :Dict = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :Optional[int] = model(**self._prepare_for_class(snake_case , snake_case ) )
snake_case_ :str = outputs.hidden_states
snake_case_ :List[Any] = getattr(
self.model_tester , """expected_num_hidden_layers""" , len(self.model_tester.depths ) + 1 )
self.assertEqual(len(snake_case ) , snake_case )
# Swinv2 has a different seq_length
snake_case_ :List[Any] = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :Optional[int] = (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] , )
snake_case_ :str = outputs.reshaped_hidden_states
self.assertEqual(len(snake_case ) , snake_case )
snake_case_, snake_case_, snake_case_, snake_case_ :Any = reshaped_hidden_states[0].shape
snake_case_ :int = (
reshaped_hidden_states[0].view(snake_case , snake_case , height * width ).permute(0 , 2 , 1 )
)
self.assertListEqual(
list(reshaped_hidden_states.shape[-2:] ) , [num_patches, self.model_tester.embed_dim] , )
def lowerCAmelCase_ ( self: Any ) -> Any:
snake_case_, snake_case_ :List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Union[str, Any] = (
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_ :Union[str, Any] = 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_ :List[str] = True
self.check_hidden_states_output(snake_case , snake_case , snake_case , snake_case )
def lowerCAmelCase_ ( self: Tuple ) -> Any:
snake_case_, snake_case_ :Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Optional[int] = 3
snake_case_ :Union[str, Any] = (
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_ :str = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :Any = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
snake_case_ :int = 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_ :Tuple = True
self.check_hidden_states_output(snake_case , snake_case , snake_case , (padded_height, padded_width) )
def lowerCAmelCase_ ( self: Any ) -> Tuple:
snake_case_ :int = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_masked_image_modeling(*snake_case )
def lowerCAmelCase_ ( self: Optional[int] ) -> Dict:
snake_case_ :Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_classification(*snake_case )
@slow
def lowerCAmelCase_ ( self: List[Any] ) -> Dict:
for model_name in SWINV2_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
snake_case_ :List[str] = SwinvaModel.from_pretrained(snake_case )
self.assertIsNotNone(snake_case )
def lowerCAmelCase_ ( self: Optional[int] ) -> List[Any]:
snake_case_, snake_case_ :str = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Optional[int] = _config_zero_init(snake_case )
for model_class in self.all_model_classes:
snake_case_ :Tuple = model_class(config=snake_case )
for name, param in model.named_parameters():
if "embeddings" not in name and "logit_scale" not in name and param.requires_grad:
self.assertIn(
((param.data.mean() * 1E9).round() / 1E9).item() , [0.0, 1.0] , msg=f"""Parameter {name} of model {model_class} seems not properly initialized""" , )
@require_vision
@require_torch
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
@cached_property
def lowerCAmelCase_ ( self: Optional[int] ) -> List[Any]:
return (
AutoImageProcessor.from_pretrained("""microsoft/swinv2-tiny-patch4-window8-256""" )
if is_vision_available()
else None
)
@slow
def lowerCAmelCase_ ( self: List[str] ) -> List[str]:
snake_case_ :Tuple = SwinvaForImageClassification.from_pretrained("""microsoft/swinv2-tiny-patch4-window8-256""" ).to(
snake_case )
snake_case_ :str = self.default_image_processor
snake_case_ :List[Any] = Image.open("""./tests/fixtures/tests_samples/COCO/000000039769.png""" )
snake_case_ :str = image_processor(images=snake_case , return_tensors="""pt""" ).to(snake_case )
# forward pass
with torch.no_grad():
snake_case_ :Tuple = model(**snake_case )
# verify the logits
snake_case_ :Dict = torch.Size((1, 1_000) )
self.assertEqual(outputs.logits.shape , snake_case )
snake_case_ :int = torch.tensor([-0.3_9_4_7, -0.4_3_0_6, 0.0_0_2_6] ).to(snake_case )
self.assertTrue(torch.allclose(outputs.logits[0, :3] , snake_case , atol=1E-4 ) )
| 66 | 0 |
from ...configuration_utils import PretrainedConfig
from ...utils import logging
A : Union[str, Any] = logging.get_logger(__name__)
A : int = {
'tiiuae/falcon-40b': 'https://huggingface.co/tiiuae/falcon-40b/resolve/main/config.json',
'tiiuae/falcon-7b': 'https://huggingface.co/tiiuae/falcon-7b/resolve/main/config.json',
}
class __A( a ):
snake_case_ = '''falcon'''
snake_case_ = ['''past_key_values''']
def __init__( self , _snake_case=65_024 , _snake_case=4_544 , _snake_case=32 , _snake_case=71 , _snake_case=1E-5 , _snake_case=0.02 , _snake_case=True , _snake_case=0.0 , _snake_case=0.0 , _snake_case=None , _snake_case=False , _snake_case=False , _snake_case=True , _snake_case=True , _snake_case=False , _snake_case=11 , _snake_case=11 , **_snake_case , ) -> List[Any]:
'''simple docstring'''
__a = vocab_size
# Backward compatibility with n_embed kwarg
__a = kwargs.pop('''n_embed''' , _snake_case )
__a = hidden_size if n_embed is None else n_embed
__a = num_hidden_layers
__a = num_attention_heads
__a = layer_norm_epsilon
__a = initializer_range
__a = use_cache
__a = hidden_dropout
__a = attention_dropout
__a = bos_token_id
__a = eos_token_id
__a = num_attention_heads if num_kv_heads is None else num_kv_heads
__a = alibi
__a = new_decoder_architecture
__a = multi_query # Ignored when new_decoder_architecture is True
__a = parallel_attn
__a = bias
super().__init__(bos_token_id=_snake_case , eos_token_id=_snake_case , **_snake_case )
@property
def SCREAMING_SNAKE_CASE_ ( self ) -> Tuple:
'''simple docstring'''
return self.hidden_size // self.num_attention_heads
@property
def SCREAMING_SNAKE_CASE_ ( self ) -> List[str]:
'''simple docstring'''
return not self.alibi | 6 |
"""simple docstring"""
import re
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Optional[int] = re.compile(
r"""^(?:0|94|\+94|0{2}94)""" r"""7(0|1|2|4|5|6|7|8)""" r"""(-| |)""" r"""\d{7}$""" )
return bool(re.search(_lowercase, _lowercase ) )
if __name__ == "__main__":
__a = "0094702343221"
print(is_sri_lankan_phone_number(phone))
| 66 | 0 |
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
lowercase_ = logging.get_logger(__name__)
lowercase_ = {
"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 A :
"""simple docstring"""
def __init__( self : Any,lowercase_ : int=None,**lowercase_ : Optional[Any] )-> Tuple:
'''simple docstring'''
logger.info('`diffusers.OnnxRuntimeModel` is experimental and might change in the future.' )
A__ = model
A__ = kwargs.get('model_save_dir',lowercase_ )
A__ = kwargs.get('latest_model_name',lowercase_ )
def __call__( self : str,**lowercase_ : Dict )-> Any:
'''simple docstring'''
A__ = {k: np.array(lowercase_ ) for k, v in kwargs.items()}
return self.model.run(lowercase_,lowercase_ )
@staticmethod
def snake_case__ ( lowercase_ : Union[str, Path],lowercase_ : Optional[Any]=None,lowercase_ : Union[str, Any]=None )-> str:
'''simple docstring'''
if provider is None:
logger.info('No onnxruntime provider specified, using CPUExecutionProvider' )
A__ = 'CPUExecutionProvider'
return ort.InferenceSession(lowercase_,providers=[provider],sess_options=lowercase_ )
def snake_case__ ( self : Dict,lowercase_ : Union[str, Path],lowercase_ : Optional[str] = None,**lowercase_ : List[str] )-> Any:
'''simple docstring'''
A__ = file_name if file_name is not None else ONNX_WEIGHTS_NAME
A__ = self.model_save_dir.joinpath(self.latest_model_name )
A__ = Path(lowercase_ ).joinpath(lowercase_ )
try:
shutil.copyfile(lowercase_,lowercase_ )
except shutil.SameFileError:
pass
# copy external weights (for models >2GB)
A__ = self.model_save_dir.joinpath(lowercase_ )
if src_path.exists():
A__ = Path(lowercase_ ).joinpath(lowercase_ )
try:
shutil.copyfile(lowercase_,lowercase_ )
except shutil.SameFileError:
pass
def snake_case__ ( self : Tuple,lowercase_ : Union[str, os.PathLike],**lowercase_ : Any,)-> List[str]:
'''simple docstring'''
if os.path.isfile(lowercase_ ):
logger.error(F'Provided path ({save_directory}) should be a directory, not a file' )
return
os.makedirs(lowercase_,exist_ok=lowercase_ )
# saving model weights/files
self._save_pretrained(lowercase_,**lowercase_ )
@classmethod
def snake_case__ ( cls : Union[str, Any],lowercase_ : Union[str, Path],lowercase_ : Optional[Union[bool, str, None]] = None,lowercase_ : Optional[Union[str, None]] = None,lowercase_ : bool = False,lowercase_ : Optional[str] = None,lowercase_ : Optional[str] = None,lowercase_ : Optional[str] = None,lowercase_ : Optional["ort.SessionOptions"] = None,**lowercase_ : int,)-> List[str]:
'''simple docstring'''
A__ = file_name if file_name is not None else ONNX_WEIGHTS_NAME
# load model from local directory
if os.path.isdir(lowercase_ ):
A__ = OnnxRuntimeModel.load_model(
os.path.join(lowercase_,lowercase_ ),provider=lowercase_,sess_options=lowercase_ )
A__ = Path(lowercase_ )
# load model from hub
else:
# download model
A__ = hf_hub_download(
repo_id=lowercase_,filename=lowercase_,use_auth_token=lowercase_,revision=lowercase_,cache_dir=lowercase_,force_download=lowercase_,)
A__ = Path(lowercase_ ).parent
A__ = Path(lowercase_ ).name
A__ = OnnxRuntimeModel.load_model(lowercase_,provider=lowercase_,sess_options=lowercase_ )
return cls(model=lowercase_,**lowercase_ )
@classmethod
def snake_case__ ( cls : str,lowercase_ : Union[str, Path],lowercase_ : bool = True,lowercase_ : Optional[str] = None,lowercase_ : Optional[str] = None,**lowercase_ : Tuple,)-> Tuple:
'''simple docstring'''
A__ = None
if len(str(lowercase_ ).split('@' ) ) == 2:
A__ , A__ = model_id.split('@' )
return cls._from_pretrained(
model_id=lowercase_,revision=lowercase_,cache_dir=lowercase_,force_download=lowercase_,use_auth_token=lowercase_,**lowercase_,)
| 7 |
"""simple docstring"""
import argparse
import json
import os
import pickle
import shutil
import numpy as np
import torch
from distiller import Distiller
from lm_seqs_dataset import LmSeqsDataset
from transformers import (
BertConfig,
BertForMaskedLM,
BertTokenizer,
DistilBertConfig,
DistilBertForMaskedLM,
DistilBertTokenizer,
GPTaConfig,
GPTaLMHeadModel,
GPTaTokenizer,
RobertaConfig,
RobertaForMaskedLM,
RobertaTokenizer,
)
from utils import git_log, init_gpu_params, logger, set_seed
__a = {
"distilbert": (DistilBertConfig, DistilBertForMaskedLM, DistilBertTokenizer),
"roberta": (RobertaConfig, RobertaForMaskedLM, RobertaTokenizer),
"bert": (BertConfig, BertForMaskedLM, BertTokenizer),
"gpt2": (GPTaConfig, GPTaLMHeadModel, GPTaTokenizer),
}
def A_ ( _lowercase ):
'''simple docstring'''
assert (args.mlm and args.alpha_mlm > 0.0) or (not args.mlm and args.alpha_mlm == 0.0)
assert (args.alpha_mlm > 0.0 and args.alpha_clm == 0.0) or (args.alpha_mlm == 0.0 and args.alpha_clm > 0.0)
if args.mlm:
assert os.path.isfile(args.token_counts )
assert (args.student_type in ["roberta", "distilbert"]) and (args.teacher_type in ["roberta", "bert"])
else:
assert (args.student_type in ["gpt2"]) and (args.teacher_type in ["gpt2"])
assert args.teacher_type == args.student_type or (
args.student_type == "distilbert" and args.teacher_type == "bert"
)
assert os.path.isfile(args.student_config )
if args.student_pretrained_weights is not None:
assert os.path.isfile(args.student_pretrained_weights )
if args.freeze_token_type_embds:
assert args.student_type in ["roberta"]
assert args.alpha_ce >= 0.0
assert args.alpha_mlm >= 0.0
assert args.alpha_clm >= 0.0
assert args.alpha_mse >= 0.0
assert args.alpha_cos >= 0.0
assert args.alpha_ce + args.alpha_mlm + args.alpha_clm + args.alpha_mse + args.alpha_cos > 0.0
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
if args.student_type == "roberta":
snake_case_ :Tuple = False
elif args.student_type == "gpt2":
snake_case_ :Union[str, Any] = False
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
if args.student_type == "roberta":
snake_case_ :List[str] = False
def A_ ( ):
'''simple docstring'''
snake_case_ :Union[str, Any] = argparse.ArgumentParser(description="""Training""" )
parser.add_argument("""--force""", action="""store_true""", help="""Overwrite dump_path if it already exists.""" )
parser.add_argument(
"""--dump_path""", type=_lowercase, required=_lowercase, help="""The output directory (log, checkpoints, parameters, etc.)""" )
parser.add_argument(
"""--data_file""", type=_lowercase, required=_lowercase, help="""The binarized file (tokenized + tokens_to_ids) and grouped by sequence.""", )
parser.add_argument(
"""--student_type""", type=_lowercase, choices=["""distilbert""", """roberta""", """gpt2"""], required=_lowercase, help="""The student type (DistilBERT, RoBERTa).""", )
parser.add_argument("""--student_config""", type=_lowercase, required=_lowercase, help="""Path to the student configuration.""" )
parser.add_argument(
"""--student_pretrained_weights""", default=_lowercase, type=_lowercase, help="""Load student initialization checkpoint.""" )
parser.add_argument(
"""--teacher_type""", choices=["""bert""", """roberta""", """gpt2"""], required=_lowercase, help="""Teacher type (BERT, RoBERTa).""" )
parser.add_argument("""--teacher_name""", type=_lowercase, required=_lowercase, help="""The teacher model.""" )
parser.add_argument("""--temperature""", default=2.0, type=_lowercase, help="""Temperature for the softmax temperature.""" )
parser.add_argument(
"""--alpha_ce""", default=0.5, type=_lowercase, help="""Linear weight for the distillation loss. Must be >=0.""" )
parser.add_argument(
"""--alpha_mlm""", default=0.0, type=_lowercase, help="""Linear weight for the MLM loss. Must be >=0. Should be used in conjunction with `mlm` flag.""", )
parser.add_argument("""--alpha_clm""", default=0.5, type=_lowercase, help="""Linear weight for the CLM loss. Must be >=0.""" )
parser.add_argument("""--alpha_mse""", default=0.0, type=_lowercase, help="""Linear weight of the MSE loss. Must be >=0.""" )
parser.add_argument(
"""--alpha_cos""", default=0.0, type=_lowercase, help="""Linear weight of the cosine embedding loss. Must be >=0.""" )
parser.add_argument(
"""--mlm""", action="""store_true""", help="""The LM step: MLM or CLM. If `mlm` is True, the MLM is used over CLM.""" )
parser.add_argument(
"""--mlm_mask_prop""", default=0.15, type=_lowercase, help="""Proportion of tokens for which we need to make a prediction.""", )
parser.add_argument("""--word_mask""", default=0.8, type=_lowercase, help="""Proportion of tokens to mask out.""" )
parser.add_argument("""--word_keep""", default=0.1, type=_lowercase, help="""Proportion of tokens to keep.""" )
parser.add_argument("""--word_rand""", default=0.1, type=_lowercase, help="""Proportion of tokens to randomly replace.""" )
parser.add_argument(
"""--mlm_smoothing""", default=0.7, type=_lowercase, help="""Smoothing parameter to emphasize more rare tokens (see XLM, similar to word2vec).""", )
parser.add_argument("""--token_counts""", type=_lowercase, help="""The token counts in the data_file for MLM.""" )
parser.add_argument(
"""--restrict_ce_to_mask""", action="""store_true""", help="""If true, compute the distillation loss only the [MLM] prediction distribution.""", )
parser.add_argument(
"""--freeze_pos_embs""", action="""store_true""", help="""Freeze positional embeddings during distillation. For student_type in ['roberta', 'gpt2'] only.""", )
parser.add_argument(
"""--freeze_token_type_embds""", action="""store_true""", help="""Freeze token type embeddings during distillation if existent. For student_type in ['roberta'] only.""", )
parser.add_argument("""--n_epoch""", type=_lowercase, default=3, help="""Number of pass on the whole dataset.""" )
parser.add_argument("""--batch_size""", type=_lowercase, default=5, help="""Batch size (for each process).""" )
parser.add_argument(
"""--group_by_size""", action="""store_false""", help="""If true, group sequences that have similar length into the same batch. Default is true.""", )
parser.add_argument(
"""--gradient_accumulation_steps""", type=_lowercase, default=50, help="""Gradient accumulation for larger training batches.""", )
parser.add_argument("""--warmup_prop""", default=0.05, type=_lowercase, help="""Linear warmup proportion.""" )
parser.add_argument("""--weight_decay""", default=0.0, type=_lowercase, help="""Weight decay if we apply some.""" )
parser.add_argument("""--learning_rate""", default=5e-4, type=_lowercase, help="""The initial learning rate for Adam.""" )
parser.add_argument("""--adam_epsilon""", default=1e-6, type=_lowercase, help="""Epsilon for Adam optimizer.""" )
parser.add_argument("""--max_grad_norm""", default=5.0, type=_lowercase, help="""Max gradient norm.""" )
parser.add_argument("""--initializer_range""", default=0.02, type=_lowercase, help="""Random initialization range.""" )
parser.add_argument(
"""--fp16""", action="""store_true""", help="""Whether to use 16-bit (mixed) precision (through NVIDIA apex) instead of 32-bit""", )
parser.add_argument(
"""--fp16_opt_level""", type=_lowercase, default="""O1""", help=(
"""For fp16: Apex AMP optimization level selected in ['O0', 'O1', 'O2', and 'O3']."""
"""See details at https://nvidia.github.io/apex/amp.html"""
), )
parser.add_argument("""--n_gpu""", type=_lowercase, default=1, help="""Number of GPUs in the node.""" )
parser.add_argument("""--local_rank""", type=_lowercase, default=-1, help="""Distributed training - Local rank""" )
parser.add_argument("""--seed""", type=_lowercase, default=56, help="""Random seed""" )
parser.add_argument("""--log_interval""", type=_lowercase, default=500, help="""Tensorboard logging interval.""" )
parser.add_argument("""--checkpoint_interval""", type=_lowercase, default=4000, help="""Checkpoint interval.""" )
snake_case_ :Tuple = parser.parse_args()
sanity_checks(_lowercase )
# ARGS #
init_gpu_params(_lowercase )
set_seed(_lowercase )
if args.is_master:
if os.path.exists(args.dump_path ):
if not args.force:
raise ValueError(
f"""Serialization dir {args.dump_path} already exists, but you have not precised wheter to overwrite"""
""" itUse `--force` if you want to overwrite it""" )
else:
shutil.rmtree(args.dump_path )
if not os.path.exists(args.dump_path ):
os.makedirs(args.dump_path )
logger.info(f"""Experiment will be dumped and logged in {args.dump_path}""" )
# SAVE PARAMS #
logger.info(f"""Param: {args}""" )
with open(os.path.join(args.dump_path, """parameters.json""" ), """w""" ) as f:
json.dump(vars(_lowercase ), _lowercase, indent=4 )
git_log(args.dump_path )
snake_case_, snake_case_, snake_case_ :Any = MODEL_CLASSES[args.student_type]
snake_case_, snake_case_, snake_case_ :int = MODEL_CLASSES[args.teacher_type]
# TOKENIZER #
snake_case_ :Any = teacher_tokenizer_class.from_pretrained(args.teacher_name )
snake_case_ :Optional[Any] = {}
for tok_name, tok_symbol in tokenizer.special_tokens_map.items():
snake_case_ :Union[str, Any] = tokenizer.all_special_tokens.index(_lowercase )
snake_case_ :Union[str, Any] = tokenizer.all_special_ids[idx]
logger.info(f"""Special tokens {special_tok_ids}""" )
snake_case_ :str = special_tok_ids
snake_case_ :Any = tokenizer.max_model_input_sizes[args.teacher_name]
# DATA LOADER #
logger.info(f"""Loading data from {args.data_file}""" )
with open(args.data_file, """rb""" ) as fp:
snake_case_ :str = pickle.load(_lowercase )
if args.mlm:
logger.info(f"""Loading token counts from {args.token_counts} (already pre-computed)""" )
with open(args.token_counts, """rb""" ) as fp:
snake_case_ :Optional[Any] = pickle.load(_lowercase )
snake_case_ :Tuple = np.maximum(_lowercase, 1 ) ** -args.mlm_smoothing
for idx in special_tok_ids.values():
snake_case_ :Optional[int] = 0.0 # do not predict special tokens
snake_case_ :int = torch.from_numpy(_lowercase )
else:
snake_case_ :List[str] = None
snake_case_ :Optional[int] = LmSeqsDataset(params=_lowercase, data=_lowercase )
logger.info("""Data loader created.""" )
# STUDENT #
logger.info(f"""Loading student config from {args.student_config}""" )
snake_case_ :List[Any] = student_config_class.from_pretrained(args.student_config )
snake_case_ :Union[str, Any] = True
if args.student_pretrained_weights is not None:
logger.info(f"""Loading pretrained weights from {args.student_pretrained_weights}""" )
snake_case_ :List[str] = student_model_class.from_pretrained(args.student_pretrained_weights, config=_lowercase )
else:
snake_case_ :Optional[int] = student_model_class(_lowercase )
if args.n_gpu > 0:
student.to(f"""cuda:{args.local_rank}""" )
logger.info("""Student loaded.""" )
# TEACHER #
snake_case_ :Dict = teacher_model_class.from_pretrained(args.teacher_name, output_hidden_states=_lowercase )
if args.n_gpu > 0:
teacher.to(f"""cuda:{args.local_rank}""" )
logger.info(f"""Teacher loaded from {args.teacher_name}.""" )
# FREEZING #
if args.freeze_pos_embs:
freeze_pos_embeddings(_lowercase, _lowercase )
if args.freeze_token_type_embds:
freeze_token_type_embeddings(_lowercase, _lowercase )
# SANITY CHECKS #
assert student.config.vocab_size == teacher.config.vocab_size
assert student.config.hidden_size == teacher.config.hidden_size
assert student.config.max_position_embeddings == teacher.config.max_position_embeddings
if args.mlm:
assert token_probs.size(0 ) == stu_architecture_config.vocab_size
# DISTILLER #
torch.cuda.empty_cache()
snake_case_ :Optional[int] = Distiller(
params=_lowercase, dataset=_lowercase, token_probs=_lowercase, student=_lowercase, teacher=_lowercase )
distiller.train()
logger.info("""Let's go get some drinks.""" )
if __name__ == "__main__":
main()
| 66 | 0 |
from statistics import mean
import numpy as np
def __SCREAMING_SNAKE_CASE (SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ):
snake_case_ = 0
# Number of processes finished
snake_case_ = 0
# Displays the finished process.
# If it is 0, the performance is completed if it is 1, before the performance.
snake_case_ = [0] * no_of_process
# List to include calculation results
snake_case_ = [0] * no_of_process
# Sort by arrival time.
snake_case_ = [burst_time[i] for i in np.argsort(SCREAMING_SNAKE_CASE__ )]
snake_case_ = [process_name[i] for i in np.argsort(SCREAMING_SNAKE_CASE__ )]
arrival_time.sort()
while no_of_process > finished_process_count:
snake_case_ = 0
while finished_process[i] == 1:
i += 1
if current_time < arrival_time[i]:
snake_case_ = arrival_time[i]
snake_case_ = 0
# Index showing the location of the process being performed
snake_case_ = 0
# Saves the current response ratio.
snake_case_ = 0
for i in range(0 , SCREAMING_SNAKE_CASE__ ):
if finished_process[i] == 0 and arrival_time[i] <= current_time:
snake_case_ = (burst_time[i] + (current_time - arrival_time[i])) / burst_time[
i
]
if response_ratio < temp:
snake_case_ = temp
snake_case_ = i
# Calculate the turn around time
snake_case_ = current_time + burst_time[loc] - arrival_time[loc]
current_time += burst_time[loc]
# Indicates that the process has been performed.
snake_case_ = 1
# Increase finished_process_count by 1
finished_process_count += 1
return turn_around_time
def __SCREAMING_SNAKE_CASE (SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ ):
snake_case_ = [0] * no_of_process
for i in range(0 , SCREAMING_SNAKE_CASE__ ):
snake_case_ = turn_around_time[i] - burst_time[i]
return waiting_time
if __name__ == "__main__":
lowerCAmelCase_ = 5
lowerCAmelCase_ = ['''A''', '''B''', '''C''', '''D''', '''E''']
lowerCAmelCase_ = [1, 2, 3, 4, 5]
lowerCAmelCase_ = [1, 2, 3, 4, 5]
lowerCAmelCase_ = calculate_turn_around_time(
process_name, arrival_time, burst_time, no_of_process
)
lowerCAmelCase_ = calculate_waiting_time(
process_name, turn_around_time, burst_time, no_of_process
)
print('''Process name \tArrival time \tBurst time \tTurn around time \tWaiting time''')
for i in range(0, no_of_process):
print(
f"""{process_name[i]}\t\t{arrival_time[i]}\t\t{burst_time[i]}\t\t"""
f"""{turn_around_time[i]}\t\t\t{waiting_time[i]}"""
)
print(f"""average waiting time : {mean(waiting_time):.5f}""")
print(f"""average turn around time : {mean(turn_around_time):.5f}""") | 8 |
"""simple docstring"""
import json
import os
import subprocess
import unittest
from ast import literal_eval
import pytest
from parameterized import parameterized, parameterized_class
from . import is_sagemaker_available
if is_sagemaker_available():
from sagemaker import Session, TrainingJobAnalytics
from sagemaker.huggingface import HuggingFace
@pytest.mark.skipif(
literal_eval(os.getenv("""TEST_SAGEMAKER""" , """False""" ) ) is not True , reason="""Skipping test because should only be run when releasing minor transformers version""" , )
@pytest.mark.usefixtures("""sm_env""" )
@parameterized_class(
[
{
"""framework""": """pytorch""",
"""script""": """run_glue_model_parallelism.py""",
"""model_name_or_path""": """roberta-large""",
"""instance_type""": """ml.p3dn.24xlarge""",
"""results""": {"""train_runtime""": 1_6_0_0, """eval_accuracy""": 0.3, """eval_loss""": 1.2},
},
{
"""framework""": """pytorch""",
"""script""": """run_glue.py""",
"""model_name_or_path""": """roberta-large""",
"""instance_type""": """ml.p3dn.24xlarge""",
"""results""": {"""train_runtime""": 1_6_0_0, """eval_accuracy""": 0.3, """eval_loss""": 1.2},
},
] )
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: Any ) -> str:
if self.framework == "pytorch":
subprocess.run(
f"""cp ./examples/pytorch/text-classification/run_glue.py {self.env.test_path}/run_glue.py""".split() , encoding="""utf-8""" , check=snake_case , )
assert hasattr(self , """env""" )
def lowerCAmelCase_ ( self: int , snake_case: Dict ) -> List[Any]:
# configuration for running training on smdistributed Model Parallel
snake_case_ :Tuple = {
"""enabled""": True,
"""processes_per_host""": 8,
}
snake_case_ :List[Any] = {
"""enabled""": True,
"""parameters""": {
"""microbatches""": 4,
"""placement_strategy""": """spread""",
"""pipeline""": """interleaved""",
"""optimize""": """speed""",
"""partitions""": 4,
"""ddp""": True,
},
}
snake_case_ :Tuple = {"""smdistributed""": {"""modelparallel""": smp_options}, """mpi""": mpi_options}
snake_case_ :Any = """trainer""" if self.script == """run_glue.py""" else """smtrainer"""
# creates estimator
return HuggingFace(
entry_point=self.script , source_dir=self.env.test_path , role=self.env.role , image_uri=self.env.image_uri , base_job_name=f"""{self.env.base_job_name}-{instance_count}-smp-{name_extension}""" , instance_count=snake_case , instance_type=self.instance_type , debugger_hook_config=snake_case , hyperparameters={
**self.env.hyperparameters,
"""model_name_or_path""": self.model_name_or_path,
"""max_steps""": 500,
} , metric_definitions=self.env.metric_definitions , distribution=snake_case , py_version="""py36""" , )
def lowerCAmelCase_ ( self: Any , snake_case: Tuple ) -> List[str]:
TrainingJobAnalytics(snake_case ).export_csv(f"""{self.env.test_path}/{job_name}_metrics.csv""" )
@parameterized.expand([(1,)] )
def lowerCAmelCase_ ( self: Dict , snake_case: Dict ) -> List[Any]:
# create estimator
snake_case_ :List[Any] = self.create_estimator(snake_case )
# run training
estimator.fit()
# result dataframe
snake_case_ :Any = TrainingJobAnalytics(estimator.latest_training_job.name ).dataframe()
# extract kpis
snake_case_ :Tuple = list(result_metrics_df[result_metrics_df.metric_name == """eval_accuracy"""]["""value"""] )
snake_case_ :Dict = list(result_metrics_df[result_metrics_df.metric_name == """eval_loss"""]["""value"""] )
# get train time from SageMaker job, this includes starting, preprocessing, stopping
snake_case_ :int = (
Session().describe_training_job(estimator.latest_training_job.name ).get("""TrainingTimeInSeconds""" , 999_999 )
)
# assert kpis
assert train_runtime <= self.results["train_runtime"]
assert all(t >= self.results["""eval_accuracy"""] for t in eval_accuracy )
assert all(t <= self.results["""eval_loss"""] for t in eval_loss )
# dump tests result into json file to share in PR
with open(f"""{estimator.latest_training_job.name}.json""" , """w""" ) as outfile:
json.dump({"""train_time""": train_runtime, """eval_accuracy""": eval_accuracy, """eval_loss""": eval_loss} , snake_case )
| 66 | 0 |
import os
from typing import List, Optional, Union
from ...tokenization_utils import PreTrainedTokenizer
from ...tokenization_utils_base import AddedToken
from ...utils import logging
__lowerCAmelCase : Any =logging.get_logger(__name__)
__lowerCAmelCase : List[str] ={'vocab_file': 'vocab.txt'}
__lowerCAmelCase : Dict ={
'vocab_file': {
'facebook/esm2_t6_8M_UR50D': 'https://huggingface.co/facebook/esm2_t6_8M_UR50D/resolve/main/vocab.txt',
'facebook/esm2_t12_35M_UR50D': 'https://huggingface.co/facebook/esm2_t12_35M_UR50D/resolve/main/vocab.txt',
},
}
__lowerCAmelCase : int ={
'facebook/esm2_t6_8M_UR50D': 1_0_2_4,
'facebook/esm2_t12_35M_UR50D': 1_0_2_4,
}
def _UpperCamelCase ( lowercase__ ):
with open(lowercase__ , '''r''' ) as f:
__SCREAMING_SNAKE_CASE : Tuple = f.read().splitlines()
return [l.strip() for l in lines]
class _lowercase ( A__ ):
'''simple docstring'''
SCREAMING_SNAKE_CASE__ : int = VOCAB_FILES_NAMES
SCREAMING_SNAKE_CASE__ : str = PRETRAINED_VOCAB_FILES_MAP
SCREAMING_SNAKE_CASE__ : Dict = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
SCREAMING_SNAKE_CASE__ : Dict = ['''input_ids''', '''attention_mask''']
def __init__( self :Dict , lowerCAmelCase__ :int , lowerCAmelCase__ :int="<unk>" , lowerCAmelCase__ :List[Any]="<cls>" , lowerCAmelCase__ :Optional[Any]="<pad>" , lowerCAmelCase__ :List[str]="<mask>" , lowerCAmelCase__ :List[Any]="<eos>" , **lowerCAmelCase__ :Optional[int] , ) -> Optional[int]:
super().__init__(**lowerCAmelCase__ )
__SCREAMING_SNAKE_CASE : List[str] = load_vocab_file(lowerCAmelCase__ )
__SCREAMING_SNAKE_CASE : Any = dict(enumerate(self.all_tokens ) )
__SCREAMING_SNAKE_CASE : Union[str, Any] = {tok: ind for ind, tok in enumerate(self.all_tokens )}
__SCREAMING_SNAKE_CASE : Tuple = unk_token
__SCREAMING_SNAKE_CASE : Tuple = cls_token
__SCREAMING_SNAKE_CASE : Dict = pad_token
__SCREAMING_SNAKE_CASE : Dict = mask_token
__SCREAMING_SNAKE_CASE : Union[str, Any] = eos_token
__SCREAMING_SNAKE_CASE : Dict = self.all_tokens
self._create_trie(self.unique_no_split_tokens )
def __magic_name__( self :str , lowerCAmelCase__ :int ) -> str:
return self._id_to_token.get(lowerCAmelCase__ , self.unk_token )
def __magic_name__( self :List[str] , lowerCAmelCase__ :str ) -> int:
return self._token_to_id.get(lowerCAmelCase__ , self._token_to_id.get(self.unk_token ) )
def __magic_name__( self :Tuple , lowerCAmelCase__ :List[str] , **lowerCAmelCase__ :Tuple ) -> Tuple:
return text.split()
def __magic_name__( self :Union[str, Any] , lowerCAmelCase__ :Optional[int]=False ) -> Optional[int]:
return len(self._id_to_token )
def __magic_name__( self :Tuple ) -> Any:
return {token: i for i, token in enumerate(self.all_tokens )}
def __magic_name__( self :str , lowerCAmelCase__ :str ) -> int:
return self._token_to_id.get(lowerCAmelCase__ , self._token_to_id.get(self.unk_token ) )
def __magic_name__( self :Optional[int] , lowerCAmelCase__ :int ) -> str:
return self._id_to_token.get(lowerCAmelCase__ , self.unk_token )
def __magic_name__( self :Tuple , lowerCAmelCase__ :List[int] , lowerCAmelCase__ :Optional[List[int]] = None ) -> List[int]:
__SCREAMING_SNAKE_CASE : str = [self.cls_token_id]
__SCREAMING_SNAKE_CASE : Optional[Any] = [self.eos_token_id] # No sep token in ESM vocabulary
if token_ids_a is None:
if self.eos_token_id is None:
return cls + token_ids_a
else:
return cls + token_ids_a + sep
elif self.eos_token_id is None:
raise ValueError('''Cannot tokenize multiple sequences when EOS token is not set!''' )
return cls + token_ids_a + sep + token_ids_a + sep # Multiple inputs always have an EOS token
def __magic_name__( self :Union[str, Any] , lowerCAmelCase__ :List , lowerCAmelCase__ :Optional[List] = None , lowerCAmelCase__ :bool = False ) -> List[int]:
if already_has_special_tokens:
if token_ids_a is not None:
raise ValueError(
'''You should not supply a second sequence if the provided sequence of '''
'''ids is already formatted with special tokens for the model.''' )
return [1 if token in self.all_special_ids else 0 for token in token_ids_a]
__SCREAMING_SNAKE_CASE : Any = [1] + ([0] * len(lowerCAmelCase__ )) + [1]
if token_ids_a is not None:
mask += [0] * len(lowerCAmelCase__ ) + [1]
return mask
def __magic_name__( self :Optional[Any] , lowerCAmelCase__ :List[str] , lowerCAmelCase__ :Union[str, Any] ) -> Optional[int]:
__SCREAMING_SNAKE_CASE : str = os.path.join(lowerCAmelCase__ , (filename_prefix + '''-''' if filename_prefix else '''''') + '''vocab.txt''' )
with open(lowerCAmelCase__ , '''w''' ) as f:
f.write('''\n'''.join(self.all_tokens ) )
return (vocab_file,)
@property
def __magic_name__( self :int ) -> int:
return self.get_vocab_size(with_added_tokens=lowerCAmelCase__ )
def __magic_name__( self :Optional[int] , lowerCAmelCase__ :Union[List[str], List[AddedToken]] , lowerCAmelCase__ :bool = False ) -> int:
return super()._add_tokens(lowerCAmelCase__ , special_tokens=lowerCAmelCase__ )
| 9 |
"""simple docstring"""
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 lowerCamelCase :
'''simple docstring'''
def __init__( self: Dict , snake_case: Optional[Any] , snake_case: Tuple=13 , snake_case: Any=32 , snake_case: Union[str, Any]=2 , snake_case: Tuple=3 , snake_case: Union[str, Any]=16 , snake_case: Union[str, Any]=[1, 2, 1] , snake_case: Optional[Any]=[2, 2, 4] , snake_case: str=2 , snake_case: List[str]=2.0 , snake_case: Optional[int]=True , snake_case: Union[str, Any]=0.0 , snake_case: Optional[int]=0.0 , snake_case: Optional[Any]=0.1 , snake_case: List[str]="gelu" , snake_case: Any=False , snake_case: Optional[Any]=True , snake_case: Optional[int]=0.0_2 , snake_case: Any=1E-5 , snake_case: Optional[int]=True , snake_case: int=None , snake_case: Any=True , snake_case: str=10 , snake_case: Optional[Any]=8 , snake_case: Union[str, Any]=["stage1", "stage2", "stage3"] , snake_case: Tuple=[1, 2, 3] , ) -> Dict:
snake_case_ :Dict = parent
snake_case_ :List[Any] = batch_size
snake_case_ :Dict = image_size
snake_case_ :Dict = patch_size
snake_case_ :Tuple = num_channels
snake_case_ :List[Any] = embed_dim
snake_case_ :List[str] = depths
snake_case_ :str = num_heads
snake_case_ :Tuple = window_size
snake_case_ :Tuple = mlp_ratio
snake_case_ :int = qkv_bias
snake_case_ :Tuple = hidden_dropout_prob
snake_case_ :Optional[Any] = attention_probs_dropout_prob
snake_case_ :Dict = drop_path_rate
snake_case_ :Any = hidden_act
snake_case_ :Any = use_absolute_embeddings
snake_case_ :int = patch_norm
snake_case_ :List[Any] = layer_norm_eps
snake_case_ :Tuple = initializer_range
snake_case_ :str = is_training
snake_case_ :int = scope
snake_case_ :Tuple = use_labels
snake_case_ :Tuple = type_sequence_label_size
snake_case_ :str = encoder_stride
snake_case_ :List[Any] = out_features
snake_case_ :str = out_indices
def lowerCAmelCase_ ( self: Tuple ) -> Dict:
snake_case_ :Optional[int] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case_ :str = None
if self.use_labels:
snake_case_ :Optional[int] = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ :Union[str, Any] = self.get_config()
return config, pixel_values, labels
def lowerCAmelCase_ ( self: int ) -> Optional[Any]:
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 lowerCAmelCase_ ( self: List[Any] , snake_case: str , snake_case: int , snake_case: List[str] ) -> Any:
snake_case_ :Dict = MaskFormerSwinModel(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Tuple = model(snake_case )
snake_case_ :Dict = ((config.image_size // config.patch_size) ** 2) // (4 ** (len(config.depths ) - 1))
snake_case_ :Any = 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 lowerCAmelCase_ ( self: Optional[Any] , snake_case: int , snake_case: List[str] , snake_case: Tuple ) -> Union[str, Any]:
snake_case_ :Any = MaskFormerSwinBackbone(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Optional[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_ :Optional[Any] = ["""stem"""]
snake_case_ :str = MaskFormerSwinBackbone(config=snake_case )
def lowerCAmelCase_ ( self: List[str] ) -> Optional[Any]:
snake_case_ :Optional[int] = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_ :str = config_and_inputs
snake_case_ :Tuple = {"""pixel_values""": pixel_values}
return config, inputs_dict
@require_torch
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : Union[str, Any] = (
(
MaskFormerSwinModel,
MaskFormerSwinBackbone,
)
if is_torch_available()
else ()
)
_A : str = {"""feature-extraction""": MaskFormerSwinModel} if is_torch_available() else {}
_A : List[str] = False
_A : Any = False
_A : Dict = False
_A : List[Any] = False
_A : Optional[int] = False
def lowerCAmelCase_ ( self: Dict ) -> Any:
snake_case_ :str = MaskFormerSwinModelTester(self )
snake_case_ :Optional[Any] = 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 lowerCAmelCase_ ( self: List[str] ) -> Optional[int]:
pass
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Dict:
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 lowerCAmelCase_ ( self: Any ) -> Tuple:
return
def lowerCAmelCase_ ( self: Any ) -> Any:
snake_case_ :List[Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> int:
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 lowerCAmelCase_ ( self: str ) -> List[str]:
pass
@unittest.skip("""Swin does not support feedforward chunking""" )
def lowerCAmelCase_ ( self: int ) -> Optional[int]:
pass
def lowerCAmelCase_ ( self: List[str] ) -> List[Any]:
snake_case_, snake_case_ :List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :str = model_class(snake_case )
self.assertIsInstance(model.get_input_embeddings() , (nn.Module) )
snake_case_ :Dict = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(snake_case , nn.Linear ) )
def lowerCAmelCase_ ( self: Tuple ) -> Dict:
snake_case_, snake_case_ :int = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :Optional[int] = model_class(snake_case )
snake_case_ :str = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case_ :str = [*signature.parameters.keys()]
snake_case_ :str = ["""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 lowerCAmelCase_ ( self: List[Any] ) -> List[Any]:
pass
@unittest.skip(reason="""MaskFormerSwin is only used as an internal backbone""" )
def lowerCAmelCase_ ( self: Dict ) -> List[Any]:
pass
def lowerCAmelCase_ ( self: Union[str, Any] , snake_case: Union[str, Any] , snake_case: int , snake_case: Any , snake_case: List[str] ) -> str:
snake_case_ :List[str] = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :List[Any] = model(**self._prepare_for_class(snake_case , snake_case ) )
snake_case_ :Any = outputs.hidden_states
snake_case_ :Optional[int] = 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_ :str = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :int = (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 lowerCAmelCase_ ( self: List[Any] ) -> Optional[int]:
snake_case_, snake_case_ :Any = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :List[Any] = (
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_ :Tuple = 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_ :List[Any] = True
self.check_hidden_states_output(snake_case , snake_case , snake_case , snake_case )
def lowerCAmelCase_ ( self: Optional[Any] ) -> Tuple:
snake_case_, snake_case_ :int = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :List[Any] = 3
snake_case_ :List[Any] = (
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_ :Any = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :Tuple = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
snake_case_ :List[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_ :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 lowerCAmelCase_ ( self: Union[str, Any] ) -> List[str]:
pass
@unittest.skip(reason="""This will be fixed once MaskFormerSwin is replaced by native Swin""" )
def lowerCAmelCase_ ( self: List[str] ) -> str:
pass
@unittest.skip(reason="""This will be fixed once MaskFormerSwin is replaced by native Swin""" )
def lowerCAmelCase_ ( self: str ) -> List[Any]:
pass
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Optional[Any]:
snake_case_, snake_case_ :Dict = self.model_tester.prepare_config_and_inputs_for_common()
def set_nan_tensor_to_zero(snake_case: str ):
snake_case_ :Optional[int] = 0
return t
def check_equivalence(snake_case: List[Any] , snake_case: Union[str, Any] , snake_case: int , snake_case: Tuple={} ):
with torch.no_grad():
snake_case_ :List[Any] = model(**snake_case , return_dict=snake_case , **snake_case )
snake_case_ :Any = model(**snake_case , return_dict=snake_case , **snake_case ).to_tuple()
def recursive_check(snake_case: List[Any] , snake_case: int ):
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_ :int = model_class(snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Any = 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 )
snake_case_ :Tuple = self._prepare_for_class(snake_case , snake_case , return_labels=snake_case )
snake_case_ :Dict = self._prepare_for_class(snake_case , snake_case , return_labels=snake_case )
check_equivalence(snake_case , snake_case , snake_case )
snake_case_ :Tuple = self._prepare_for_class(snake_case , snake_case )
snake_case_ :Any = self._prepare_for_class(snake_case , snake_case )
check_equivalence(snake_case , snake_case , snake_case , {"""output_hidden_states""": True} )
snake_case_ :Dict = 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 lowerCamelCase ( unittest.TestCase , _lowerCAmelCase ):
'''simple docstring'''
_A : int = (MaskFormerSwinBackbone,) if is_torch_available() else ()
_A : Tuple = MaskFormerSwinConfig
def lowerCAmelCase_ ( self: List[str] ) -> Optional[int]:
snake_case_ :Optional[Any] = MaskFormerSwinModelTester(self )
def lowerCAmelCase_ ( self: int ) -> Optional[int]:
snake_case_, snake_case_ :Any = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Tuple = inputs_dict["""pixel_values"""].shape[0]
for backbone_class in self.all_model_classes:
snake_case_ :List[str] = backbone_class(snake_case )
backbone.to(snake_case )
backbone.eval()
snake_case_ :List[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_ :Union[str, Any] = 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_ :List[Any] = 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_ :List[Any] = backbone(**snake_case , output_attentions=snake_case )
self.assertIsNotNone(outputs.attentions )
| 66 | 0 |
def lowerCAmelCase_ ( __a ) -> List[Any]:
"""simple docstring"""
if collection == []:
return []
# get some information about the collection
lowerCamelCase__: List[Any] =len(__a )
lowerCamelCase__: List[str] =max(__a )
lowerCamelCase__: Dict =min(__a )
# create the counting array
lowerCamelCase__: Tuple =coll_max + 1 - coll_min
lowerCamelCase__: Optional[int] =[0] * counting_arr_length
# count how much a number appears in the collection
for number in collection:
counting_arr[number - coll_min] += 1
# sum each position with it's predecessors. now, counting_arr[i] tells
# us how many elements <= i has in the collection
for i in range(1 , __a ):
lowerCamelCase__: int =counting_arr[i] + counting_arr[i - 1]
# create the output collection
lowerCamelCase__: Dict =[0] * coll_len
# place the elements in the output, respecting the original order (stable
# sort) from end to begin, updating counting_arr
for i in reversed(range(0 , __a ) ):
lowerCamelCase__: int =collection[i]
counting_arr[collection[i] - coll_min] -= 1
return ordered
def lowerCAmelCase_ ( __a ) -> str:
"""simple docstring"""
return "".join([chr(__a ) for i in counting_sort([ord(__a ) for c in string] )] )
if __name__ == "__main__":
# Test string sort
assert counting_sort_string("thisisthestring") == "eghhiiinrsssttt"
__A = input("Enter numbers separated by a comma:\n").strip()
__A = [int(item) for item in user_input.split(",")]
print(counting_sort(unsorted))
| 10 |
"""simple docstring"""
import gc
import math
import unittest
import torch
from diffusers import UNetaDModel
from diffusers.utils import floats_tensor, logging, slow, torch_all_close, torch_device
from diffusers.utils.testing_utils import enable_full_determinism
from .test_modeling_common import ModelTesterMixin, UNetTesterMixin
__a = logging.get_logger(__name__)
enable_full_determinism()
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[Any] = UNetaDModel
_A : Union[str, Any] = """sample"""
@property
def lowerCAmelCase_ ( self: str ) -> Tuple:
snake_case_ :List[str] = 4
snake_case_ :Tuple = 3
snake_case_ :Optional[Any] = (32, 32)
snake_case_ :str = floats_tensor((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :Union[str, Any] = torch.tensor([10] ).to(snake_case )
return {"sample": noise, "timestep": time_step}
@property
def lowerCAmelCase_ ( self: List[str] ) -> Dict:
return (3, 32, 32)
@property
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
return (3, 32, 32)
def lowerCAmelCase_ ( self: Optional[int] ) -> Dict:
snake_case_ :Any = {
"""block_out_channels""": (32, 64),
"""down_block_types""": ("""DownBlock2D""", """AttnDownBlock2D"""),
"""up_block_types""": ("""AttnUpBlock2D""", """UpBlock2D"""),
"""attention_head_dim""": 3,
"""out_channels""": 3,
"""in_channels""": 3,
"""layers_per_block""": 2,
"""sample_size""": 32,
}
snake_case_ :Tuple = self.dummy_input
return init_dict, inputs_dict
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[str] = UNetaDModel
_A : Union[str, Any] = """sample"""
@property
def lowerCAmelCase_ ( self: str ) -> str:
snake_case_ :List[str] = 4
snake_case_ :Optional[int] = 4
snake_case_ :int = (32, 32)
snake_case_ :Any = floats_tensor((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :List[Any] = torch.tensor([10] ).to(snake_case )
return {"sample": noise, "timestep": time_step}
@property
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
return (4, 32, 32)
@property
def lowerCAmelCase_ ( self: List[Any] ) -> int:
return (4, 32, 32)
def lowerCAmelCase_ ( self: Union[str, Any] ) -> List[Any]:
snake_case_ :Dict = {
"""sample_size""": 32,
"""in_channels""": 4,
"""out_channels""": 4,
"""layers_per_block""": 2,
"""block_out_channels""": (32, 64),
"""attention_head_dim""": 32,
"""down_block_types""": ("""DownBlock2D""", """DownBlock2D"""),
"""up_block_types""": ("""UpBlock2D""", """UpBlock2D"""),
}
snake_case_ :List[str] = self.dummy_input
return init_dict, inputs_dict
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[Any]:
snake_case_, snake_case_ :List[str] = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case )
self.assertIsNotNone(snake_case )
self.assertEqual(len(loading_info["""missing_keys"""] ) , 0 )
model.to(snake_case )
snake_case_ :List[str] = model(**self.dummy_input ).sample
assert image is not None, "Make sure output is not None"
@unittest.skipIf(torch_device != """cuda""" , """This test is supposed to run on GPU""" )
def lowerCAmelCase_ ( self: Tuple ) -> Dict:
snake_case_, snake_case_ :Union[str, Any] = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case )
model.to(snake_case )
snake_case_ :Union[str, Any] = model(**self.dummy_input ).sample
assert image is not None, "Make sure output is not None"
@unittest.skipIf(torch_device != """cuda""" , """This test is supposed to run on GPU""" )
def lowerCAmelCase_ ( self: str ) -> Any:
# by defautl model loading will use accelerate as `low_cpu_mem_usage=True`
snake_case_, snake_case_ :List[str] = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case )
model_accelerate.to(snake_case )
model_accelerate.eval()
snake_case_ :List[Any] = torch.randn(
1 , model_accelerate.config.in_channels , model_accelerate.config.sample_size , model_accelerate.config.sample_size , generator=torch.manual_seed(0 ) , )
snake_case_ :int = noise.to(snake_case )
snake_case_ :str = torch.tensor([10] * noise.shape[0] ).to(snake_case )
snake_case_ :Optional[int] = model_accelerate(snake_case , snake_case )["""sample"""]
# two models don't need to stay in the device at the same time
del model_accelerate
torch.cuda.empty_cache()
gc.collect()
snake_case_, snake_case_ :str = UNetaDModel.from_pretrained(
"""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case , low_cpu_mem_usage=snake_case )
model_normal_load.to(snake_case )
model_normal_load.eval()
snake_case_ :int = model_normal_load(snake_case , snake_case )["""sample"""]
assert torch_all_close(snake_case , snake_case , rtol=1E-3 )
def lowerCAmelCase_ ( self: Tuple ) -> Any:
snake_case_ :Tuple = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" )
model.eval()
model.to(snake_case )
snake_case_ :Optional[int] = torch.randn(
1 , model.config.in_channels , model.config.sample_size , model.config.sample_size , generator=torch.manual_seed(0 ) , )
snake_case_ :int = noise.to(snake_case )
snake_case_ :List[Any] = torch.tensor([10] * noise.shape[0] ).to(snake_case )
with torch.no_grad():
snake_case_ :Union[str, Any] = model(snake_case , snake_case ).sample
snake_case_ :Optional[int] = output[0, -1, -3:, -3:].flatten().cpu()
# fmt: off
snake_case_ :Dict = torch.tensor([-1_3.3_2_5_8, -2_0.1_1_0_0, -1_5.9_8_7_3, -1_7.6_6_1_7, -2_3.0_5_9_6, -1_7.9_4_1_9, -1_3.3_6_7_5, -1_6.1_8_8_9, -1_2.3_8_0_0] )
# fmt: on
self.assertTrue(torch_all_close(snake_case , snake_case , rtol=1E-3 ) )
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[Any] = UNetaDModel
_A : List[Any] = """sample"""
@property
def lowerCAmelCase_ ( self: Union[str, Any] , snake_case: int=(32, 32) ) -> Tuple:
snake_case_ :Union[str, Any] = 4
snake_case_ :Any = 3
snake_case_ :int = floats_tensor((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :Any = torch.tensor(batch_size * [10] ).to(dtype=torch.intaa , device=snake_case )
return {"sample": noise, "timestep": time_step}
@property
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Any:
return (3, 32, 32)
@property
def lowerCAmelCase_ ( self: int ) -> Tuple:
return (3, 32, 32)
def lowerCAmelCase_ ( self: List[str] ) -> Tuple:
snake_case_ :List[Any] = {
"""block_out_channels""": [32, 64, 64, 64],
"""in_channels""": 3,
"""layers_per_block""": 1,
"""out_channels""": 3,
"""time_embedding_type""": """fourier""",
"""norm_eps""": 1E-6,
"""mid_block_scale_factor""": math.sqrt(2.0 ),
"""norm_num_groups""": None,
"""down_block_types""": [
"""SkipDownBlock2D""",
"""AttnSkipDownBlock2D""",
"""SkipDownBlock2D""",
"""SkipDownBlock2D""",
],
"""up_block_types""": [
"""SkipUpBlock2D""",
"""SkipUpBlock2D""",
"""AttnSkipUpBlock2D""",
"""SkipUpBlock2D""",
],
}
snake_case_ :int = self.dummy_input
return init_dict, inputs_dict
@slow
def lowerCAmelCase_ ( self: Optional[Any] ) -> List[Any]:
snake_case_, snake_case_ :List[Any] = UNetaDModel.from_pretrained("""google/ncsnpp-celebahq-256""" , output_loading_info=snake_case )
self.assertIsNotNone(snake_case )
self.assertEqual(len(loading_info["""missing_keys"""] ) , 0 )
model.to(snake_case )
snake_case_ :Any = self.dummy_input
snake_case_ :int = floats_tensor((4, 3) + (256, 256) ).to(snake_case )
snake_case_ :int = noise
snake_case_ :int = model(**snake_case )
assert image is not None, "Make sure output is not None"
@slow
def lowerCAmelCase_ ( self: str ) -> Dict:
snake_case_ :Dict = UNetaDModel.from_pretrained("""google/ncsnpp-celebahq-256""" )
model.to(snake_case )
snake_case_ :List[str] = 4
snake_case_ :Optional[int] = 3
snake_case_ :List[str] = (256, 256)
snake_case_ :Tuple = torch.ones((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :str = torch.tensor(batch_size * [1E-4] ).to(snake_case )
with torch.no_grad():
snake_case_ :Dict = model(snake_case , snake_case ).sample
snake_case_ :int = output[0, -3:, -3:, -1].flatten().cpu()
# fmt: off
snake_case_ :Optional[Any] = torch.tensor([-4_8_4_2.8_6_9_1, -6_4_9_9.6_6_3_1, -3_8_0_0.1_9_5_3, -7_9_7_8.2_6_8_6, -1_0_9_8_0.7_1_2_9, -2_0_0_2_8.8_5_3_5, 8_1_4_8.2_8_2_2, 2_3_4_2.2_9_0_5, 5_6_7.7_6_0_8] )
# fmt: on
self.assertTrue(torch_all_close(snake_case , snake_case , rtol=1E-2 ) )
def lowerCAmelCase_ ( self: List[str] ) -> List[Any]:
snake_case_ :Optional[Any] = UNetaDModel.from_pretrained("""fusing/ncsnpp-ffhq-ve-dummy-update""" )
model.to(snake_case )
snake_case_ :Optional[int] = 4
snake_case_ :Optional[Any] = 3
snake_case_ :Optional[Any] = (32, 32)
snake_case_ :Dict = torch.ones((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :Any = torch.tensor(batch_size * [1E-4] ).to(snake_case )
with torch.no_grad():
snake_case_ :str = model(snake_case , snake_case ).sample
snake_case_ :int = output[0, -3:, -3:, -1].flatten().cpu()
# fmt: off
snake_case_ :int = torch.tensor([-0.0_3_2_5, -0.0_9_0_0, -0.0_8_6_9, -0.0_3_3_2, -0.0_7_2_5, -0.0_2_7_0, -0.0_1_0_1, 0.0_2_2_7, 0.0_2_5_6] )
# fmt: on
self.assertTrue(torch_all_close(snake_case , snake_case , rtol=1E-2 ) )
def lowerCAmelCase_ ( self: Dict ) -> Optional[Any]:
# not required for this model
pass
| 66 | 0 |
from ..utils import (
OptionalDependencyNotAvailable,
is_flax_available,
is_scipy_available,
is_torch_available,
is_torchsde_available,
)
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from ..utils.dummy_pt_objects import * # noqa F403
else:
from .scheduling_consistency_models import CMStochasticIterativeScheduler
from .scheduling_ddim import DDIMScheduler
from .scheduling_ddim_inverse import DDIMInverseScheduler
from .scheduling_ddim_parallel import DDIMParallelScheduler
from .scheduling_ddpm import DDPMScheduler
from .scheduling_ddpm_parallel import DDPMParallelScheduler
from .scheduling_deis_multistep import DEISMultistepScheduler
from .scheduling_dpmsolver_multistep import DPMSolverMultistepScheduler
from .scheduling_dpmsolver_multistep_inverse import DPMSolverMultistepInverseScheduler
from .scheduling_dpmsolver_singlestep import DPMSolverSinglestepScheduler
from .scheduling_euler_ancestral_discrete import EulerAncestralDiscreteScheduler
from .scheduling_euler_discrete import EulerDiscreteScheduler
from .scheduling_heun_discrete import HeunDiscreteScheduler
from .scheduling_ipndm import IPNDMScheduler
from .scheduling_k_dpm_2_ancestral_discrete import KDPMaAncestralDiscreteScheduler
from .scheduling_k_dpm_2_discrete import KDPMaDiscreteScheduler
from .scheduling_karras_ve import KarrasVeScheduler
from .scheduling_pndm import PNDMScheduler
from .scheduling_repaint import RePaintScheduler
from .scheduling_sde_ve import ScoreSdeVeScheduler
from .scheduling_sde_vp import ScoreSdeVpScheduler
from .scheduling_unclip import UnCLIPScheduler
from .scheduling_unipc_multistep import UniPCMultistepScheduler
from .scheduling_utils import KarrasDiffusionSchedulers, SchedulerMixin
from .scheduling_vq_diffusion import VQDiffusionScheduler
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
from ..utils.dummy_flax_objects import * # noqa F403
else:
from .scheduling_ddim_flax import FlaxDDIMScheduler
from .scheduling_ddpm_flax import FlaxDDPMScheduler
from .scheduling_dpmsolver_multistep_flax import FlaxDPMSolverMultistepScheduler
from .scheduling_karras_ve_flax import FlaxKarrasVeScheduler
from .scheduling_lms_discrete_flax import FlaxLMSDiscreteScheduler
from .scheduling_pndm_flax import FlaxPNDMScheduler
from .scheduling_sde_ve_flax import FlaxScoreSdeVeScheduler
from .scheduling_utils_flax import (
FlaxKarrasDiffusionSchedulers,
FlaxSchedulerMixin,
FlaxSchedulerOutput,
broadcast_to_shape_from_left,
)
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 .scheduling_lms_discrete 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 .scheduling_dpmsolver_sde import DPMSolverSDEScheduler
| 11 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available, is_vision_available
__a = {
"configuration_mask2former": [
"MASK2FORMER_PRETRAINED_CONFIG_ARCHIVE_MAP",
"Mask2FormerConfig",
],
}
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["Mask2FormerImageProcessor"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"MASK2FORMER_PRETRAINED_MODEL_ARCHIVE_LIST",
"Mask2FormerForUniversalSegmentation",
"Mask2FormerModel",
"Mask2FormerPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_maskaformer import MASK2FORMER_PRETRAINED_CONFIG_ARCHIVE_MAP, MaskaFormerConfig
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .image_processing_maskaformer import MaskaFormerImageProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_maskaformer import (
MASK2FORMER_PRETRAINED_MODEL_ARCHIVE_LIST,
MaskaFormerForUniversalSegmentation,
MaskaFormerModel,
MaskaFormerPreTrainedModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure)
| 66 | 0 |
import importlib
import torch
import yaml
from omegaconf import OmegaConf
from taming.models.vqgan import VQModel
def lowerCamelCase__ ( A__ : Dict , A__ : Optional[int]=False ):
'''simple docstring'''
__lowerCamelCase = OmegaConf.load(A__ )
if display:
print(yaml.dump(OmegaConf.to_container(A__ ) ) )
return config
def lowerCamelCase__ ( A__ : Optional[int] , A__ : Union[str, Any]=None , A__ : Any=None ):
'''simple docstring'''
if conf_path is None:
__lowerCamelCase = """./model_checkpoints/vqgan_only.yaml"""
__lowerCamelCase = load_config(A__ , display=A__ )
__lowerCamelCase = VQModel(**config.model.params )
if ckpt_path is None:
__lowerCamelCase = """./model_checkpoints/vqgan_only.pt"""
__lowerCamelCase = torch.load(A__ , map_location=A__ )
if ".ckpt" in ckpt_path:
__lowerCamelCase = sd["""state_dict"""]
model.load_state_dict(A__ , strict=A__ )
model.to(A__ )
del sd
return model
def lowerCamelCase__ ( A__ : Optional[Any] , A__ : List[Any] ):
'''simple docstring'''
__lowerCamelCase, __lowerCamelCase, __lowerCamelCase = model.encode(A__ )
print(f'VQGAN --- {model.__class__.__name__}: latent shape: {z.shape[2:]}' )
__lowerCamelCase = model.decode(A__ )
return xrec
def lowerCamelCase__ ( A__ : Tuple , A__ : List[Any]=False ):
'''simple docstring'''
__lowerCamelCase, __lowerCamelCase = string.rsplit(""".""" , 1 )
if reload:
__lowerCamelCase = importlib.import_module(A__ )
importlib.reload(A__ )
return getattr(importlib.import_module(A__ , package=A__ ) , cls )
def lowerCamelCase__ ( A__ : Optional[int] ):
'''simple docstring'''
if "target" not in config:
raise KeyError("""Expected key `target` to instantiate.""" )
return get_obj_from_str(config["""target"""] )(**config.get("""params""" , {} ) )
def lowerCamelCase__ ( A__ : Optional[Any] , A__ : Optional[Any] , A__ : Dict=True , A__ : int=True ):
'''simple docstring'''
__lowerCamelCase = instantiate_from_config(A__ )
if sd is not None:
model.load_state_dict(A__ )
if gpu:
model.cuda()
if eval_mode:
model.eval()
return {"model": model}
def lowerCamelCase__ ( A__ : List[Any] , A__ : str , A__ : Dict , A__ : List[Any] ):
'''simple docstring'''
if ckpt:
__lowerCamelCase = torch.load(A__ , map_location="""cpu""" )
__lowerCamelCase = pl_sd["""global_step"""]
print(f'loaded model from global step {global_step}.' )
else:
__lowerCamelCase = {"""state_dict""": None}
__lowerCamelCase = None
__lowerCamelCase = load_model_from_config(config.model , pl_sd["""state_dict"""] , gpu=A__ , eval_mode=A__ )["""model"""]
return model, global_step
| 12 |
"""simple docstring"""
import gc
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import (
AutoencoderKL,
DDIMScheduler,
StableDiffusionSAGPipeline,
UNetaDConditionModel,
)
from diffusers.utils import slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu
from ..pipeline_params import TEXT_TO_IMAGE_BATCH_PARAMS, TEXT_TO_IMAGE_IMAGE_PARAMS, TEXT_TO_IMAGE_PARAMS
from ..test_pipelines_common import PipelineLatentTesterMixin, PipelineTesterMixin
enable_full_determinism()
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : str = StableDiffusionSAGPipeline
_A : Optional[Any] = TEXT_TO_IMAGE_PARAMS
_A : Any = TEXT_TO_IMAGE_BATCH_PARAMS
_A : Tuple = TEXT_TO_IMAGE_IMAGE_PARAMS
_A : Tuple = TEXT_TO_IMAGE_IMAGE_PARAMS
_A : List[str] = False
def lowerCAmelCase_ ( self: Optional[Any] ) -> str:
torch.manual_seed(0 )
snake_case_ :Any = UNetaDConditionModel(
block_out_channels=(32, 64) , layers_per_block=2 , sample_size=32 , in_channels=4 , out_channels=4 , down_block_types=("""DownBlock2D""", """CrossAttnDownBlock2D""") , up_block_types=("""CrossAttnUpBlock2D""", """UpBlock2D""") , cross_attention_dim=32 , )
snake_case_ :Any = DDIMScheduler(
beta_start=0.0_0_0_8_5 , beta_end=0.0_1_2 , beta_schedule="""scaled_linear""" , clip_sample=snake_case , set_alpha_to_one=snake_case , )
torch.manual_seed(0 )
snake_case_ :Optional[int] = AutoencoderKL(
block_out_channels=[32, 64] , in_channels=3 , out_channels=3 , down_block_types=["""DownEncoderBlock2D""", """DownEncoderBlock2D"""] , up_block_types=["""UpDecoderBlock2D""", """UpDecoderBlock2D"""] , latent_channels=4 , )
torch.manual_seed(0 )
snake_case_ :Union[str, Any] = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1E-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1_000 , )
snake_case_ :Tuple = CLIPTextModel(snake_case )
snake_case_ :str = CLIPTokenizer.from_pretrained("""hf-internal-testing/tiny-random-clip""" )
snake_case_ :Dict = {
"""unet""": unet,
"""scheduler""": scheduler,
"""vae""": vae,
"""text_encoder""": text_encoder,
"""tokenizer""": tokenizer,
"""safety_checker""": None,
"""feature_extractor""": None,
}
return components
def lowerCAmelCase_ ( self: List[str] , snake_case: Tuple , snake_case: List[str]=0 ) -> str:
if str(snake_case ).startswith("""mps""" ):
snake_case_ :Tuple = torch.manual_seed(snake_case )
else:
snake_case_ :Optional[int] = torch.Generator(device=snake_case ).manual_seed(snake_case )
snake_case_ :Any = {
"""prompt""": """.""",
"""generator""": generator,
"""num_inference_steps""": 2,
"""guidance_scale""": 1.0,
"""sag_scale""": 1.0,
"""output_type""": """numpy""",
}
return inputs
def lowerCAmelCase_ ( self: Optional[int] ) -> str:
super().test_inference_batch_single_identical(expected_max_diff=3E-3 )
@slow
@require_torch_gpu
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: int ) -> str:
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def lowerCAmelCase_ ( self: int ) -> List[str]:
snake_case_ :Any = StableDiffusionSAGPipeline.from_pretrained("""CompVis/stable-diffusion-v1-4""" )
snake_case_ :int = sag_pipe.to(snake_case )
sag_pipe.set_progress_bar_config(disable=snake_case )
snake_case_ :Union[str, Any] = """."""
snake_case_ :str = torch.manual_seed(0 )
snake_case_ :str = sag_pipe(
[prompt] , generator=snake_case , guidance_scale=7.5 , sag_scale=1.0 , num_inference_steps=20 , output_type="""np""" )
snake_case_ :List[Any] = output.images
snake_case_ :Tuple = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
snake_case_ :List[Any] = np.array([0.1_5_6_8, 0.1_7_3_8, 0.1_6_9_5, 0.1_6_9_3, 0.1_5_0_7, 0.1_7_0_5, 0.1_5_4_7, 0.1_7_5_1, 0.1_9_4_9] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 5E-2
def lowerCAmelCase_ ( self: Dict ) -> str:
snake_case_ :Tuple = StableDiffusionSAGPipeline.from_pretrained("""stabilityai/stable-diffusion-2-1-base""" )
snake_case_ :Optional[int] = sag_pipe.to(snake_case )
sag_pipe.set_progress_bar_config(disable=snake_case )
snake_case_ :Tuple = """."""
snake_case_ :Union[str, Any] = torch.manual_seed(0 )
snake_case_ :Tuple = sag_pipe(
[prompt] , generator=snake_case , guidance_scale=7.5 , sag_scale=1.0 , num_inference_steps=20 , output_type="""np""" )
snake_case_ :Optional[int] = output.images
snake_case_ :Optional[Any] = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
snake_case_ :Tuple = np.array([0.3_4_5_9, 0.2_8_7_6, 0.2_5_3_7, 0.3_0_0_2, 0.2_6_7_1, 0.2_1_6_0, 0.3_0_2_6, 0.2_2_6_2, 0.2_3_7_1] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 5E-2
def lowerCAmelCase_ ( self: List[str] ) -> List[str]:
snake_case_ :Optional[int] = StableDiffusionSAGPipeline.from_pretrained("""stabilityai/stable-diffusion-2-1-base""" )
snake_case_ :int = sag_pipe.to(snake_case )
sag_pipe.set_progress_bar_config(disable=snake_case )
snake_case_ :Tuple = """."""
snake_case_ :Optional[int] = torch.manual_seed(0 )
snake_case_ :List[str] = sag_pipe(
[prompt] , width=768 , height=512 , generator=snake_case , guidance_scale=7.5 , sag_scale=1.0 , num_inference_steps=20 , output_type="""np""" , )
snake_case_ :Optional[Any] = output.images
assert image.shape == (1, 512, 768, 3)
| 66 | 0 |
# Algorithm for the pigeonhole sorting
def A_ ( _UpperCAmelCase ):
SCREAMING_SNAKE_CASE_: Optional[Any] = min(_UpperCAmelCase ) # min() finds the minimum value
SCREAMING_SNAKE_CASE_: Tuple = max(_UpperCAmelCase ) # max() finds the maximum value
SCREAMING_SNAKE_CASE_: str = max_val - min_val + 1 # size is difference of max and min values plus one
# list of pigeonholes of size equal to the variable size
SCREAMING_SNAKE_CASE_: List[Any] = [0] * size
# Populate the pigeonholes.
for x in a:
assert isinstance(_UpperCAmelCase , _UpperCAmelCase ), "integers only please"
holes[x - min_val] += 1
# Putting the elements back into the array in an order.
SCREAMING_SNAKE_CASE_: int = 0
for count in range(_UpperCAmelCase ):
while holes[count] > 0:
holes[count] -= 1
SCREAMING_SNAKE_CASE_: str = count + min_val
i += 1
def A_ ( ):
SCREAMING_SNAKE_CASE_: List[str] = [8, 3, 2, 7, 4, 6, 8]
pigeonhole_sort(_UpperCAmelCase )
print("Sorted order is:" , " ".join(_UpperCAmelCase ) )
if __name__ == "__main__":
main()
| 13 |
"""simple docstring"""
from __future__ import annotations
from collections import Counter
from random import random
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Tuple ) -> Optional[Any]:
snake_case_ :Optional[int] = {}
def lowerCAmelCase_ ( self: Dict , snake_case: str ) -> None:
snake_case_ :str = {}
def lowerCAmelCase_ ( self: Optional[int] , snake_case: str , snake_case: str , snake_case: float ) -> None:
if nodea not in self.connections:
self.add_node(snake_case )
if nodea not in self.connections:
self.add_node(snake_case )
snake_case_ :Dict = probability
def lowerCAmelCase_ ( self: List[Any] ) -> list[str]:
return list(self.connections )
def lowerCAmelCase_ ( self: Any , snake_case: str ) -> str:
snake_case_ :Optional[Any] = 0
snake_case_ :List[str] = random()
for dest in self.connections[node]:
current_probability += self.connections[node][dest]
if current_probability > random_value:
return dest
return ""
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :List[str] = MarkovChainGraphUndirectedUnweighted()
for nodea, nodea, probability in transitions:
graph.add_transition_probability(_lowercase, _lowercase, _lowercase )
snake_case_ :int = Counter(graph.get_nodes() )
snake_case_ :Optional[Any] = start
for _ in range(_lowercase ):
snake_case_ :Tuple = graph.transition(_lowercase )
visited[node] += 1
return visited
if __name__ == "__main__":
import doctest
doctest.testmod()
| 66 | 0 |
from __future__ import annotations
_lowerCamelCase : Union[str, Any] = 10
def SCREAMING_SNAKE_CASE ( lowercase_ ) -> list[int]:
"""simple docstring"""
A__ = 1
A__ = max(lowercase_ )
while placement <= max_digit:
# declare and initialize empty buckets
A__ = [[] for _ in range(lowercase_ )]
# split list_of_ints between the buckets
for i in list_of_ints:
A__ = int((i / placement) % RADIX )
buckets[tmp].append(lowercase_ )
# put each buckets' contents into list_of_ints
A__ = 0
for b in range(lowercase_ ):
for i in buckets[b]:
A__ = i
a += 1
# move to next
placement *= RADIX
return list_of_ints
if __name__ == "__main__":
import doctest
doctest.testmod()
| 14 |
"""simple docstring"""
import argparse
import collections
import os
import re
import tempfile
import pandas as pd
from datasets import Dataset
from huggingface_hub import hf_hub_download, upload_folder
from transformers.utils import direct_transformers_import
# All paths are set with the intent you should run this script from the root of the repo with the command
# python utils/update_metadata.py
__a = "src/transformers"
# This is to make sure the transformers module imported is the one in the repo.
__a = direct_transformers_import(TRANSFORMERS_PATH)
# Regexes that match TF/Flax/PT model names.
__a = re.compile(r"TF(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
__a = re.compile(r"Flax(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
# Will match any TF or Flax model too so need to be in an else branch afterthe two previous regexes.
__a = re.compile(r"(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
# Fill this with tuples (pipeline_tag, model_mapping, auto_model)
__a = [
("pretraining", "MODEL_FOR_PRETRAINING_MAPPING_NAMES", "AutoModelForPreTraining"),
("feature-extraction", "MODEL_MAPPING_NAMES", "AutoModel"),
("audio-classification", "MODEL_FOR_AUDIO_CLASSIFICATION_MAPPING_NAMES", "AutoModelForAudioClassification"),
("text-generation", "MODEL_FOR_CAUSAL_LM_MAPPING_NAMES", "AutoModelForCausalLM"),
("automatic-speech-recognition", "MODEL_FOR_CTC_MAPPING_NAMES", "AutoModelForCTC"),
("image-classification", "MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES", "AutoModelForImageClassification"),
("image-segmentation", "MODEL_FOR_IMAGE_SEGMENTATION_MAPPING_NAMES", "AutoModelForImageSegmentation"),
("fill-mask", "MODEL_FOR_MASKED_LM_MAPPING_NAMES", "AutoModelForMaskedLM"),
("object-detection", "MODEL_FOR_OBJECT_DETECTION_MAPPING_NAMES", "AutoModelForObjectDetection"),
(
"zero-shot-object-detection",
"MODEL_FOR_ZERO_SHOT_OBJECT_DETECTION_MAPPING_NAMES",
"AutoModelForZeroShotObjectDetection",
),
("question-answering", "MODEL_FOR_QUESTION_ANSWERING_MAPPING_NAMES", "AutoModelForQuestionAnswering"),
("text2text-generation", "MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING_NAMES", "AutoModelForSeq2SeqLM"),
("text-classification", "MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES", "AutoModelForSequenceClassification"),
("automatic-speech-recognition", "MODEL_FOR_SPEECH_SEQ_2_SEQ_MAPPING_NAMES", "AutoModelForSpeechSeq2Seq"),
(
"table-question-answering",
"MODEL_FOR_TABLE_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForTableQuestionAnswering",
),
("token-classification", "MODEL_FOR_TOKEN_CLASSIFICATION_MAPPING_NAMES", "AutoModelForTokenClassification"),
("multiple-choice", "MODEL_FOR_MULTIPLE_CHOICE_MAPPING_NAMES", "AutoModelForMultipleChoice"),
(
"next-sentence-prediction",
"MODEL_FOR_NEXT_SENTENCE_PREDICTION_MAPPING_NAMES",
"AutoModelForNextSentencePrediction",
),
(
"audio-frame-classification",
"MODEL_FOR_AUDIO_FRAME_CLASSIFICATION_MAPPING_NAMES",
"AutoModelForAudioFrameClassification",
),
("audio-xvector", "MODEL_FOR_AUDIO_XVECTOR_MAPPING_NAMES", "AutoModelForAudioXVector"),
(
"document-question-answering",
"MODEL_FOR_DOCUMENT_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForDocumentQuestionAnswering",
),
(
"visual-question-answering",
"MODEL_FOR_VISUAL_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForVisualQuestionAnswering",
),
("image-to-text", "MODEL_FOR_FOR_VISION_2_SEQ_MAPPING_NAMES", "AutoModelForVision2Seq"),
(
"zero-shot-image-classification",
"MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING_NAMES",
"AutoModelForZeroShotImageClassification",
),
("depth-estimation", "MODEL_FOR_DEPTH_ESTIMATION_MAPPING_NAMES", "AutoModelForDepthEstimation"),
("video-classification", "MODEL_FOR_VIDEO_CLASSIFICATION_MAPPING_NAMES", "AutoModelForVideoClassification"),
("mask-generation", "MODEL_FOR_MASK_GENERATION_MAPPING_NAMES", "AutoModelForMaskGeneration"),
]
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Any = re.finditer(""".+?(?:(?<=[a-z])(?=[A-Z])|(?<=[A-Z])(?=[A-Z][a-z])|$)""", _lowercase )
return [m.group(0 ) for m in matches]
def A_ ( ):
'''simple docstring'''
snake_case_ :int = transformers_module.models.auto.configuration_auto.CONFIG_MAPPING_NAMES
snake_case_ :Dict = {
config.replace("""Config""", """""" ): model_type for model_type, config in config_maping_names.items()
}
# Dictionaries flagging if each model prefix has a backend in PT/TF/Flax.
snake_case_ :Optional[Any] = collections.defaultdict(_lowercase )
snake_case_ :int = collections.defaultdict(_lowercase )
snake_case_ :List[str] = collections.defaultdict(_lowercase )
# Let's lookup through all transformers object (once) and find if models are supported by a given backend.
for attr_name in dir(_lowercase ):
snake_case_ :int = None
if _re_tf_models.match(_lowercase ) is not None:
snake_case_ :int = tf_models
snake_case_ :List[str] = _re_tf_models.match(_lowercase ).groups()[0]
elif _re_flax_models.match(_lowercase ) is not None:
snake_case_ :List[Any] = flax_models
snake_case_ :Any = _re_flax_models.match(_lowercase ).groups()[0]
elif _re_pt_models.match(_lowercase ) is not None:
snake_case_ :Optional[Any] = pt_models
snake_case_ :int = _re_pt_models.match(_lowercase ).groups()[0]
if lookup_dict is not None:
while len(_lowercase ) > 0:
if attr_name in model_prefix_to_model_type:
snake_case_ :Optional[int] = True
break
# Try again after removing the last word in the name
snake_case_ :Optional[Any] = """""".join(camel_case_split(_lowercase )[:-1] )
snake_case_ :Optional[int] = set(list(pt_models.keys() ) + list(tf_models.keys() ) + list(flax_models.keys() ) )
snake_case_ :Optional[Any] = list(_lowercase )
all_models.sort()
snake_case_ :Optional[int] = {"""model_type""": all_models}
snake_case_ :Optional[int] = [pt_models[t] for t in all_models]
snake_case_ :Any = [tf_models[t] for t in all_models]
snake_case_ :Dict = [flax_models[t] for t in all_models]
# Now let's use the auto-mapping names to make sure
snake_case_ :Dict = {}
for t in all_models:
if t in transformers_module.models.auto.processing_auto.PROCESSOR_MAPPING_NAMES:
snake_case_ :Optional[Any] = """AutoProcessor"""
elif t in transformers_module.models.auto.tokenization_auto.TOKENIZER_MAPPING_NAMES:
snake_case_ :Tuple = """AutoTokenizer"""
elif t in transformers_module.models.auto.feature_extraction_auto.FEATURE_EXTRACTOR_MAPPING_NAMES:
snake_case_ :Tuple = """AutoFeatureExtractor"""
else:
# Default to AutoTokenizer if a model has nothing, for backward compatibility.
snake_case_ :str = """AutoTokenizer"""
snake_case_ :int = [processors[t] for t in all_models]
return pd.DataFrame(_lowercase )
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :List[Any] = [
transformers_module.models.auto.modeling_auto,
transformers_module.models.auto.modeling_tf_auto,
transformers_module.models.auto.modeling_flax_auto,
]
for pipeline_tag, model_mapping, auto_class in PIPELINE_TAGS_AND_AUTO_MODELS:
snake_case_ :Optional[int] = [model_mapping, f"""TF_{model_mapping}""", f"""FLAX_{model_mapping}"""]
snake_case_ :List[str] = [auto_class, f"""TF_{auto_class}""", f"""Flax_{auto_class}"""]
# Loop through all three frameworks
for module, cls, mapping in zip(_lowercase, _lowercase, _lowercase ):
# The type of pipeline may not exist in this framework
if not hasattr(_lowercase, _lowercase ):
continue
# First extract all model_names
snake_case_ :Tuple = []
for name in getattr(_lowercase, _lowercase ).values():
if isinstance(_lowercase, _lowercase ):
model_names.append(_lowercase )
else:
model_names.extend(list(_lowercase ) )
# Add pipeline tag and auto model class for those models
table.update({model_name: (pipeline_tag, cls) for model_name in model_names} )
return table
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :List[Any] = get_frameworks_table()
snake_case_ :str = Dataset.from_pandas(_lowercase )
snake_case_ :List[Any] = hf_hub_download(
"""huggingface/transformers-metadata""", """pipeline_tags.json""", repo_type="""dataset""", token=_lowercase )
snake_case_ :List[str] = Dataset.from_json(_lowercase )
snake_case_ :int = {
tags_dataset[i]["""model_class"""]: (tags_dataset[i]["""pipeline_tag"""], tags_dataset[i]["""auto_class"""])
for i in range(len(_lowercase ) )
}
snake_case_ :Optional[int] = update_pipeline_and_auto_class_table(_lowercase )
# Sort the model classes to avoid some nondeterministic updates to create false update commits.
snake_case_ :Tuple = sorted(table.keys() )
snake_case_ :Tuple = pd.DataFrame(
{
"""model_class""": model_classes,
"""pipeline_tag""": [table[m][0] for m in model_classes],
"""auto_class""": [table[m][1] for m in model_classes],
} )
snake_case_ :Union[str, Any] = Dataset.from_pandas(_lowercase )
with tempfile.TemporaryDirectory() as tmp_dir:
frameworks_dataset.to_json(os.path.join(_lowercase, """frameworks.json""" ) )
tags_dataset.to_json(os.path.join(_lowercase, """pipeline_tags.json""" ) )
if commit_sha is not None:
snake_case_ :Union[str, Any] = (
f"""Update with commit {commit_sha}\n\nSee: """
f"""https://github.com/huggingface/transformers/commit/{commit_sha}"""
)
else:
snake_case_ :List[Any] = """Update"""
upload_folder(
repo_id="""huggingface/transformers-metadata""", folder_path=_lowercase, repo_type="""dataset""", token=_lowercase, commit_message=_lowercase, )
def A_ ( ):
'''simple docstring'''
snake_case_ :List[Any] = {tag: cls for tag, _, cls in PIPELINE_TAGS_AND_AUTO_MODELS}
snake_case_ :Dict = transformers_module.pipelines.SUPPORTED_TASKS
snake_case_ :List[str] = []
for key in pipeline_tasks:
if key not in in_table:
snake_case_ :int = pipeline_tasks[key]["""pt"""]
if isinstance(_lowercase, (list, tuple) ):
snake_case_ :Any = model[0]
snake_case_ :str = model.__name__
if model not in in_table.values():
missing.append(_lowercase )
if len(_lowercase ) > 0:
snake_case_ :Optional[int] = """, """.join(_lowercase )
raise ValueError(
"""The following pipeline tags are not present in the `PIPELINE_TAGS_AND_AUTO_MODELS` constant inside """
f"""`utils/update_metadata.py`: {msg}. Please add them!""" )
if __name__ == "__main__":
__a = argparse.ArgumentParser()
parser.add_argument("--token", type=str, help="The token to use to push to the transformers-metadata dataset.")
parser.add_argument("--commit_sha", type=str, help="The sha of the commit going with this update.")
parser.add_argument("--check-only", action="store_true", help="Activate to just check all pipelines are present.")
__a = parser.parse_args()
if args.check_only:
check_pipeline_tags()
else:
update_metadata(args.token, args.commit_sha)
| 66 | 0 |
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_flax_available,
is_tf_available,
is_torch_available,
)
SCREAMING_SNAKE_CASE :str = {
'configuration_vision_text_dual_encoder': ['VisionTextDualEncoderConfig'],
'processing_vision_text_dual_encoder': ['VisionTextDualEncoderProcessor'],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
SCREAMING_SNAKE_CASE :Dict = ['VisionTextDualEncoderModel']
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
SCREAMING_SNAKE_CASE :Any = ['FlaxVisionTextDualEncoderModel']
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
SCREAMING_SNAKE_CASE :List[str] = ['TFVisionTextDualEncoderModel']
if TYPE_CHECKING:
from .configuration_vision_text_dual_encoder import VisionTextDualEncoderConfig
from .processing_vision_text_dual_encoder import VisionTextDualEncoderProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_vision_text_dual_encoder import VisionTextDualEncoderModel
try:
if not is_flax_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_flax_vision_text_dual_encoder import FlaxVisionTextDualEncoderModel
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_vision_text_dual_encoder import TFVisionTextDualEncoderModel
else:
import sys
SCREAMING_SNAKE_CASE :str = _LazyModule(__name__, globals()['__file__'], _import_structure)
| 15 |
"""simple docstring"""
import argparse
import glob
import logging
import os
from argparse import Namespace
from importlib import import_module
import numpy as np
import torch
from lightning_base import BaseTransformer, add_generic_args, generic_train
from seqeval.metrics import accuracy_score, fa_score, precision_score, recall_score
from torch.nn import CrossEntropyLoss
from torch.utils.data import DataLoader, TensorDataset
from utils_ner import TokenClassificationTask
__a = logging.getLogger(__name__)
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
_A : Union[str, Any] = """token-classification"""
def __init__( self: Any , snake_case: Tuple ) -> List[Any]:
if type(snake_case ) == dict:
snake_case_ :Optional[int] = Namespace(**snake_case )
snake_case_ :Optional[int] = import_module("""tasks""" )
try:
snake_case_ :Any = getattr(snake_case , hparams.task_type )
snake_case_ :TokenClassificationTask = token_classification_task_clazz()
except AttributeError:
raise ValueError(
f"""Task {hparams.task_type} needs to be defined as a TokenClassificationTask subclass in {module}. """
f"""Available tasks classes are: {TokenClassificationTask.__subclasses__()}""" )
snake_case_ :Any = self.token_classification_task.get_labels(hparams.labels )
snake_case_ :str = CrossEntropyLoss().ignore_index
super().__init__(snake_case , len(self.labels ) , self.mode )
def lowerCAmelCase_ ( self: Dict , **snake_case: List[Any] ) -> Any:
return self.model(**snake_case )
def lowerCAmelCase_ ( self: str , snake_case: Tuple , snake_case: List[Any] ) -> Optional[int]:
snake_case_ :List[str] = {"""input_ids""": batch[0], """attention_mask""": batch[1], """labels""": batch[3]}
if self.config.model_type != "distilbert":
snake_case_ :List[str] = (
batch[2] if self.config.model_type in ["""bert""", """xlnet"""] else None
) # XLM and RoBERTa don"t use token_type_ids
snake_case_ :Optional[Any] = self(**snake_case )
snake_case_ :List[str] = outputs[0]
# tensorboard_logs = {"loss": loss, "rate": self.lr_scheduler.get_last_lr()[-1]}
return {"loss": loss}
def lowerCAmelCase_ ( self: int ) -> Dict:
snake_case_ :List[Any] = self.hparams
for mode in ["train", "dev", "test"]:
snake_case_ :Optional[int] = self._feature_file(snake_case )
if os.path.exists(snake_case ) and not args.overwrite_cache:
logger.info("""Loading features from cached file %s""" , snake_case )
snake_case_ :Optional[int] = torch.load(snake_case )
else:
logger.info("""Creating features from dataset file at %s""" , args.data_dir )
snake_case_ :Optional[int] = self.token_classification_task.read_examples_from_file(args.data_dir , snake_case )
snake_case_ :Any = self.token_classification_task.convert_examples_to_features(
snake_case , self.labels , args.max_seq_length , self.tokenizer , cls_token_at_end=bool(self.config.model_type in ["""xlnet"""] ) , cls_token=self.tokenizer.cls_token , cls_token_segment_id=2 if self.config.model_type in ["""xlnet"""] else 0 , sep_token=self.tokenizer.sep_token , sep_token_extra=snake_case , pad_on_left=bool(self.config.model_type in ["""xlnet"""] ) , pad_token=self.tokenizer.pad_token_id , pad_token_segment_id=self.tokenizer.pad_token_type_id , pad_token_label_id=self.pad_token_label_id , )
logger.info("""Saving features into cached file %s""" , snake_case )
torch.save(snake_case , snake_case )
def lowerCAmelCase_ ( self: Optional[int] , snake_case: int , snake_case: int , snake_case: bool = False ) -> DataLoader:
snake_case_ :int = self._feature_file(snake_case )
logger.info("""Loading features from cached file %s""" , snake_case )
snake_case_ :str = torch.load(snake_case )
snake_case_ :Dict = torch.tensor([f.input_ids for f in features] , dtype=torch.long )
snake_case_ :str = torch.tensor([f.attention_mask for f in features] , dtype=torch.long )
if features[0].token_type_ids is not None:
snake_case_ :List[Any] = torch.tensor([f.token_type_ids for f in features] , dtype=torch.long )
else:
snake_case_ :List[str] = torch.tensor([0 for f in features] , dtype=torch.long )
# HACK(we will not use this anymore soon)
snake_case_ :Any = torch.tensor([f.label_ids for f in features] , dtype=torch.long )
return DataLoader(
TensorDataset(snake_case , snake_case , snake_case , snake_case ) , batch_size=snake_case )
def lowerCAmelCase_ ( self: List[str] , snake_case: Dict , snake_case: Union[str, Any] ) -> List[str]:
"""Compute validation""" ""
snake_case_ :List[str] = {"""input_ids""": batch[0], """attention_mask""": batch[1], """labels""": batch[3]}
if self.config.model_type != "distilbert":
snake_case_ :Dict = (
batch[2] if self.config.model_type in ["""bert""", """xlnet"""] else None
) # XLM and RoBERTa don"t use token_type_ids
snake_case_ :Dict = self(**snake_case )
snake_case_, snake_case_ :Dict = outputs[:2]
snake_case_ :Union[str, Any] = logits.detach().cpu().numpy()
snake_case_ :List[Any] = inputs["""labels"""].detach().cpu().numpy()
return {"val_loss": tmp_eval_loss.detach().cpu(), "pred": preds, "target": out_label_ids}
def lowerCAmelCase_ ( self: List[Any] , snake_case: int ) -> Tuple:
snake_case_ :Union[str, Any] = torch.stack([x["""val_loss"""] for x in outputs] ).mean()
snake_case_ :Tuple = np.concatenate([x["""pred"""] for x in outputs] , axis=0 )
snake_case_ :Tuple = np.argmax(snake_case , axis=2 )
snake_case_ :List[str] = np.concatenate([x["""target"""] for x in outputs] , axis=0 )
snake_case_ :Optional[Any] = dict(enumerate(self.labels ) )
snake_case_ :Dict = [[] for _ in range(out_label_ids.shape[0] )]
snake_case_ :Dict = [[] for _ in range(out_label_ids.shape[0] )]
for i in range(out_label_ids.shape[0] ):
for j in range(out_label_ids.shape[1] ):
if out_label_ids[i, j] != self.pad_token_label_id:
out_label_list[i].append(label_map[out_label_ids[i][j]] )
preds_list[i].append(label_map[preds[i][j]] )
snake_case_ :str = {
"""val_loss""": val_loss_mean,
"""accuracy_score""": accuracy_score(snake_case , snake_case ),
"""precision""": precision_score(snake_case , snake_case ),
"""recall""": recall_score(snake_case , snake_case ),
"""f1""": fa_score(snake_case , snake_case ),
}
snake_case_ :List[Any] = dict(results.items() )
snake_case_ :Union[str, Any] = results
return ret, preds_list, out_label_list
def lowerCAmelCase_ ( self: Optional[Any] , snake_case: Dict ) -> Optional[Any]:
# when stable
snake_case_, snake_case_, snake_case_ :Tuple = self._eval_end(snake_case )
snake_case_ :str = ret["""log"""]
return {"val_loss": logs["val_loss"], "log": logs, "progress_bar": logs}
def lowerCAmelCase_ ( self: Tuple , snake_case: Optional[int] ) -> Any:
# updating to test_epoch_end instead of deprecated test_end
snake_case_, snake_case_, snake_case_ :Any = self._eval_end(snake_case )
# Converting to the dict required by pl
# https://github.com/PyTorchLightning/pytorch-lightning/blob/master/\
# pytorch_lightning/trainer/logging.py#L139
snake_case_ :Optional[int] = ret["""log"""]
# `val_loss` is the key returned by `self._eval_end()` but actually refers to `test_loss`
return {"avg_test_loss": logs["val_loss"], "log": logs, "progress_bar": logs}
@staticmethod
def lowerCAmelCase_ ( snake_case: Any , snake_case: int ) -> Dict:
# Add NER specific options
BaseTransformer.add_model_specific_args(snake_case , snake_case )
parser.add_argument(
"""--task_type""" , default="""NER""" , type=snake_case , help="""Task type to fine tune in training (e.g. NER, POS, etc)""" )
parser.add_argument(
"""--max_seq_length""" , default=128 , type=snake_case , help=(
"""The maximum total input sequence length after tokenization. Sequences longer """
"""than this will be truncated, sequences shorter will be padded."""
) , )
parser.add_argument(
"""--labels""" , default="""""" , type=snake_case , help="""Path to a file containing all labels. If not specified, CoNLL-2003 labels are used.""" , )
parser.add_argument(
"""--gpus""" , default=0 , type=snake_case , help="""The number of GPUs allocated for this, it is by default 0 meaning none""" , )
parser.add_argument(
"""--overwrite_cache""" , action="""store_true""" , help="""Overwrite the cached training and evaluation sets""" )
return parser
if __name__ == "__main__":
__a = argparse.ArgumentParser()
add_generic_args(parser, os.getcwd())
__a = NERTransformer.add_model_specific_args(parser, os.getcwd())
__a = parser.parse_args()
__a = NERTransformer(args)
__a = generic_train(model, args)
if args.do_predict:
# See https://github.com/huggingface/transformers/issues/3159
# pl use this default format to create a checkpoint:
# https://github.com/PyTorchLightning/pytorch-lightning/blob/master\
# /pytorch_lightning/callbacks/model_checkpoint.py#L322
__a = sorted(glob.glob(os.path.join(args.output_dir, "checkpoint-epoch=*.ckpt"), recursive=True))
__a = model.load_from_checkpoint(checkpoints[-1])
trainer.test(model)
| 66 | 0 |
"""simple docstring"""
import os
import socket
from contextlib import contextmanager
import torch
from ..commands.config.default import write_basic_config # noqa: F401
from ..state import PartialState
from .dataclasses import DistributedType
from .imports import is_deepspeed_available, is_tpu_available
from .transformer_engine import convert_model
from .versions import is_torch_version
if is_deepspeed_available():
from deepspeed import DeepSpeedEngine
if is_tpu_available(check_device=False):
import torch_xla.core.xla_model as xm
def __UpperCAmelCase ( __lowerCamelCase ) -> Any:
if is_torch_version('''<''' , '''2.0.0''' ) or not hasattr(__lowerCamelCase , '''_dynamo''' ):
return False
return isinstance(__lowerCamelCase , torch._dynamo.eval_frame.OptimizedModule )
def __UpperCAmelCase ( __lowerCamelCase , __lowerCamelCase = True ) -> Dict:
lowercase__ : Optional[Any] = (torch.nn.parallel.DistributedDataParallel, torch.nn.DataParallel)
lowercase__ : Any = is_compiled_module(__lowerCamelCase )
if is_compiled:
lowercase__ : Dict = model
lowercase__ : Tuple = model._orig_mod
if is_deepspeed_available():
options += (DeepSpeedEngine,)
while isinstance(__lowerCamelCase , __lowerCamelCase ):
lowercase__ : int = model.module
if not keep_fpaa_wrapper:
lowercase__ : str = getattr(__lowerCamelCase , '''forward''' )
lowercase__ : List[str] = model.__dict__.pop('''_original_forward''' , __lowerCamelCase )
if original_forward is not None:
while hasattr(__lowerCamelCase , '''__wrapped__''' ):
lowercase__ : List[Any] = forward.__wrapped__
if forward == original_forward:
break
lowercase__ : Optional[Any] = forward
if getattr(__lowerCamelCase , '''_converted_to_transformer_engine''' , __lowerCamelCase ):
convert_model(__lowerCamelCase , to_transformer_engine=__lowerCamelCase )
if is_compiled:
lowercase__ : Optional[Any] = model
lowercase__ : Tuple = compiled_model
return model
def __UpperCAmelCase ( ) -> str:
PartialState().wait_for_everyone()
def __UpperCAmelCase ( __lowerCamelCase , __lowerCamelCase ) -> Union[str, Any]:
if PartialState().distributed_type == DistributedType.TPU:
xm.save(__lowerCamelCase , __lowerCamelCase )
elif PartialState().local_process_index == 0:
torch.save(__lowerCamelCase , __lowerCamelCase )
@contextmanager
def __UpperCAmelCase ( **__lowerCamelCase ) -> Optional[Any]:
for key, value in kwargs.items():
lowercase__ : Tuple = str(__lowerCamelCase )
yield
for key in kwargs:
if key.upper() in os.environ:
del os.environ[key.upper()]
def __UpperCAmelCase ( __lowerCamelCase ) -> List[Any]:
if not hasattr(__lowerCamelCase , '''__qualname__''' ) and not hasattr(__lowerCamelCase , '''__name__''' ):
lowercase__ : Tuple = getattr(__lowerCamelCase , '''__class__''' , __lowerCamelCase )
if hasattr(__lowerCamelCase , '''__qualname__''' ):
return obj.__qualname__
if hasattr(__lowerCamelCase , '''__name__''' ):
return obj.__name__
return str(__lowerCamelCase )
def __UpperCAmelCase ( __lowerCamelCase , __lowerCamelCase ) -> str:
for key, value in source.items():
if isinstance(__lowerCamelCase , __lowerCamelCase ):
lowercase__ : List[str] = destination.setdefault(__lowerCamelCase , {} )
merge_dicts(__lowerCamelCase , __lowerCamelCase )
else:
lowercase__ : Any = value
return destination
def __UpperCAmelCase ( __lowerCamelCase = None ) -> bool:
if port is None:
lowercase__ : Optional[Any] = 2_95_00
with socket.socket(socket.AF_INET , socket.SOCK_STREAM ) as s:
return s.connect_ex(('''localhost''', port) ) == 0
| 16 |
"""simple docstring"""
from math import factorial
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Optional[int] , snake_case: Dict , snake_case: int ) -> Tuple:
snake_case_ :List[Any] = real
if isinstance(snake_case , snake_case ):
snake_case_ :Tuple = [1] * rank
else:
snake_case_ :Optional[Any] = rank
def __repr__( self: List[str] ) -> Tuple:
return (
f"""{self.real}+"""
f"""{'+'.join(str(snake_case )+'E'+str(n+1 )for n,dual in enumerate(self.duals ) )}"""
)
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
snake_case_ :Any = self.duals.copy()
while cur[-1] == 0:
cur.pop(-1 )
return Dual(self.real , snake_case )
def __add__( self: Optional[int] , snake_case: Dict ) -> List[str]:
if not isinstance(snake_case , snake_case ):
return Dual(self.real + other , self.duals )
snake_case_ :List[Any] = self.duals.copy()
snake_case_ :Tuple = other.duals.copy()
if len(snake_case ) > len(snake_case ):
o_dual.extend([1] * (len(snake_case ) - len(snake_case )) )
elif len(snake_case ) < len(snake_case ):
s_dual.extend([1] * (len(snake_case ) - len(snake_case )) )
snake_case_ :Dict = []
for i in range(len(snake_case ) ):
new_duals.append(s_dual[i] + o_dual[i] )
return Dual(self.real + other.real , snake_case )
_A : str = __add__
def __sub__( self: Tuple , snake_case: Union[str, Any] ) -> Tuple:
return self + other * -1
def __mul__( self: str , snake_case: Tuple ) -> Optional[Any]:
if not isinstance(snake_case , snake_case ):
snake_case_ :Dict = []
for i in self.duals:
new_duals.append(i * other )
return Dual(self.real * other , snake_case )
snake_case_ :int = [0] * (len(self.duals ) + len(other.duals ) + 1)
for i, item in enumerate(self.duals ):
for j, jtem in enumerate(other.duals ):
new_duals[i + j + 1] += item * jtem
for k in range(len(self.duals ) ):
new_duals[k] += self.duals[k] * other.real
for index in range(len(other.duals ) ):
new_duals[index] += other.duals[index] * self.real
return Dual(self.real * other.real , snake_case )
_A : int = __mul__
def __truediv__( self: List[str] , snake_case: List[str] ) -> List[str]:
if not isinstance(snake_case , snake_case ):
snake_case_ :Optional[Any] = []
for i in self.duals:
new_duals.append(i / other )
return Dual(self.real / other , snake_case )
raise ValueError
def __floordiv__( self: int , snake_case: List[Any] ) -> Any:
if not isinstance(snake_case , snake_case ):
snake_case_ :Optional[int] = []
for i in self.duals:
new_duals.append(i // other )
return Dual(self.real // other , snake_case )
raise ValueError
def __pow__( self: Optional[Any] , snake_case: Optional[int] ) -> List[Any]:
if n < 0 or isinstance(snake_case , snake_case ):
raise ValueError("""power must be a positive integer""" )
if n == 0:
return 1
if n == 1:
return self
snake_case_ :str = self
for _ in range(n - 1 ):
x *= self
return x
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
if not callable(_lowercase ):
raise ValueError("""differentiate() requires a function as input for func""" )
if not isinstance(_lowercase, (float, int) ):
raise ValueError("""differentiate() requires a float as input for position""" )
if not isinstance(_lowercase, _lowercase ):
raise ValueError("""differentiate() requires an int as input for order""" )
snake_case_ :Optional[Any] = Dual(_lowercase, 1 )
snake_case_ :List[Any] = func(_lowercase )
if order == 0:
return result.real
return result.duals[order - 1] * factorial(_lowercase )
if __name__ == "__main__":
import doctest
doctest.testmod()
def A_ ( _lowercase ):
'''simple docstring'''
return y**2 * y**4
print(differentiate(f, 9, 2))
| 66 | 0 |
"""simple docstring"""
from ...configuration_utils import PretrainedConfig
from ...utils import logging
_a = logging.get_logger(__name__)
_a = {
'facebook/nllb-moe-54B': 'https://huggingface.co/facebook/nllb-moe-54b/resolve/main/config.json',
}
class _lowerCAmelCase ( lowercase ):
"""simple docstring"""
__UpperCAmelCase : Any = "nllb-moe"
__UpperCAmelCase : str = ["past_key_values"]
__UpperCAmelCase : Any = {"num_attention_heads": "encoder_attention_heads", "hidden_size": "d_model"}
def __init__( self : List[Any], UpperCAmelCase__ : str=1_2_8_1_1_2, UpperCAmelCase__ : int=1_0_2_4, UpperCAmelCase__ : Optional[Any]=1_2, UpperCAmelCase__ : List[str]=4_0_9_6, UpperCAmelCase__ : Dict=1_6, UpperCAmelCase__ : Tuple=1_2, UpperCAmelCase__ : Union[str, Any]=4_0_9_6, UpperCAmelCase__ : int=1_6, UpperCAmelCase__ : Optional[int]=0.05, UpperCAmelCase__ : Union[str, Any]=0.05, UpperCAmelCase__ : Optional[int]=True, UpperCAmelCase__ : str=True, UpperCAmelCase__ : Tuple="relu", UpperCAmelCase__ : Optional[int]=1_0_2_4, UpperCAmelCase__ : str=0.1, UpperCAmelCase__ : Any=0.1, UpperCAmelCase__ : Optional[int]=0.0, UpperCAmelCase__ : Dict=0.02, UpperCAmelCase__ : str=2, UpperCAmelCase__ : List[Any]=True, UpperCAmelCase__ : int=False, UpperCAmelCase__ : Dict="float32", UpperCAmelCase__ : List[str]=False, UpperCAmelCase__ : Any=1_2_8, UpperCAmelCase__ : Any=6_4, UpperCAmelCase__ : str=4, UpperCAmelCase__ : Any=4, UpperCAmelCase__ : int=0.001, UpperCAmelCase__ : Optional[int]=0.001, UpperCAmelCase__ : Optional[Any]="all", UpperCAmelCase__ : Optional[int]=False, UpperCAmelCase__ : str=False, UpperCAmelCase__ : Dict=1.0, UpperCAmelCase__ : List[str]=0.2, UpperCAmelCase__ : Any=1, UpperCAmelCase__ : Optional[Any]=0, UpperCAmelCase__ : Dict=2, UpperCAmelCase__ : List[str]=False, **UpperCAmelCase__ : int, ):
__lowercase = vocab_size
__lowercase = max_position_embeddings
__lowercase = d_model
__lowercase = encoder_ffn_dim
__lowercase = encoder_layers
__lowercase = encoder_attention_heads
__lowercase = decoder_ffn_dim
__lowercase = decoder_layers
__lowercase = decoder_attention_heads
__lowercase = dropout
__lowercase = attention_dropout
__lowercase = activation_dropout
__lowercase = activation_function
__lowercase = init_std
__lowercase = encoder_layerdrop
__lowercase = decoder_layerdrop
__lowercase = use_cache
__lowercase = encoder_layers
__lowercase = scale_embedding # scale factor will be sqrt(d_model) if True
__lowercase = router_z_loss_coef
__lowercase = router_aux_loss_coef
__lowercase = decoder_sparse_step
__lowercase = encoder_sparse_step
__lowercase = num_experts
__lowercase = expert_capacity
__lowercase = router_bias
if router_dtype not in ["float32", "float16", "bfloat16"]:
raise ValueError(F"""`router_dtype` must be one of 'float32', 'float16' or 'bfloat16', got {router_dtype}""" )
__lowercase = router_dtype
__lowercase = router_ignore_padding_tokens
__lowercase = batch_prioritized_routing
__lowercase = second_expert_policy
__lowercase = normalize_router_prob_before_dropping
__lowercase = moe_eval_capacity_token_fraction
__lowercase = moe_token_dropout
__lowercase = output_router_logits
super().__init__(
pad_token_id=UpperCAmelCase__, bos_token_id=UpperCAmelCase__, eos_token_id=UpperCAmelCase__, is_encoder_decoder=UpperCAmelCase__, decoder_start_token_id=UpperCAmelCase__, **UpperCAmelCase__, )
| 17 |
"""simple docstring"""
from __future__ import annotations
__a = 10
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Union[str, Any] = 1
snake_case_ :List[str] = max(_lowercase )
while placement <= max_digit:
# declare and initialize empty buckets
snake_case_ :list[list] = [[] for _ in range(_lowercase )]
# split list_of_ints between the buckets
for i in list_of_ints:
snake_case_ :Any = int((i / placement) % RADIX )
buckets[tmp].append(_lowercase )
# put each buckets' contents into list_of_ints
snake_case_ :Optional[Any] = 0
for b in range(_lowercase ):
for i in buckets[b]:
snake_case_ :Union[str, Any] = i
a += 1
# move to next
placement *= RADIX
return list_of_ints
if __name__ == "__main__":
import doctest
doctest.testmod()
| 66 | 0 |
def _snake_case ( lowerCAmelCase : int = 1_0_0 ):
"""simple docstring"""
SCREAMING_SNAKE_CASE_ : Optional[int] = n * (n + 1) * (2 * n + 1) / 6
SCREAMING_SNAKE_CASE_ : int = (n * (n + 1) / 2) ** 2
return int(square_of_sum - sum_of_squares )
if __name__ == "__main__":
print(f'''{solution() = }''')
| 18 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_sentencepiece_available,
is_tokenizers_available,
is_torch_available,
)
__a = {"configuration_reformer": ["REFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP", "ReformerConfig"]}
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["ReformerTokenizer"]
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["ReformerTokenizerFast"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"REFORMER_PRETRAINED_MODEL_ARCHIVE_LIST",
"ReformerAttention",
"ReformerForMaskedLM",
"ReformerForQuestionAnswering",
"ReformerForSequenceClassification",
"ReformerLayer",
"ReformerModel",
"ReformerModelWithLMHead",
"ReformerPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_reformer import REFORMER_PRETRAINED_CONFIG_ARCHIVE_MAP, ReformerConfig
try:
if not is_sentencepiece_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_reformer import ReformerTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_reformer_fast import ReformerTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_reformer import (
REFORMER_PRETRAINED_MODEL_ARCHIVE_LIST,
ReformerAttention,
ReformerForMaskedLM,
ReformerForQuestionAnswering,
ReformerForSequenceClassification,
ReformerLayer,
ReformerModel,
ReformerModelWithLMHead,
ReformerPreTrainedModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 | 0 |
import json
from typing import List, Optional, Tuple
from tokenizers import normalizers
from ....tokenization_utils_fast import PreTrainedTokenizerFast
from ....utils import logging
from .tokenization_retribert import RetriBertTokenizer
__A =logging.get_logger(__name__)
__A ={'''vocab_file''': '''vocab.txt''', '''tokenizer_file''': '''tokenizer.json'''}
__A ={
'''vocab_file''': {
'''yjernite/retribert-base-uncased''': (
'''https://huggingface.co/yjernite/retribert-base-uncased/resolve/main/vocab.txt'''
),
},
'''tokenizer_file''': {
'''yjernite/retribert-base-uncased''': (
'''https://huggingface.co/yjernite/retribert-base-uncased/resolve/main/tokenizer.json'''
),
},
}
__A ={
'''yjernite/retribert-base-uncased''': 5_1_2,
}
__A ={
'''yjernite/retribert-base-uncased''': {'''do_lower_case''': True},
}
class _SCREAMING_SNAKE_CASE ( snake_case_ ):
lowerCAmelCase__ = VOCAB_FILES_NAMES
lowerCAmelCase__ = PRETRAINED_VOCAB_FILES_MAP
lowerCAmelCase__ = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
lowerCAmelCase__ = PRETRAINED_INIT_CONFIGURATION
lowerCAmelCase__ = RetriBertTokenizer
lowerCAmelCase__ = ['input_ids', 'attention_mask']
def __init__( self , lowercase=None , lowercase=None , lowercase=True , lowercase="[UNK]" , lowercase="[SEP]" , lowercase="[PAD]" , lowercase="[CLS]" , lowercase="[MASK]" , lowercase=True , lowercase=None , **lowercase , ) -> List[Any]:
super().__init__(
lowercase , tokenizer_file=lowercase , do_lower_case=lowercase , unk_token=lowercase , sep_token=lowercase , pad_token=lowercase , cls_token=lowercase , mask_token=lowercase , tokenize_chinese_chars=lowercase , strip_accents=lowercase , **lowercase , )
lowerCamelCase_ = json.loads(self.backend_tokenizer.normalizer.__getstate__() )
if (
normalizer_state.get("lowercase" , lowercase ) != do_lower_case
or normalizer_state.get("strip_accents" , lowercase ) != strip_accents
or normalizer_state.get("handle_chinese_chars" , lowercase ) != tokenize_chinese_chars
):
lowerCamelCase_ = getattr(lowercase , normalizer_state.pop("type" ) )
lowerCamelCase_ = do_lower_case
lowerCamelCase_ = strip_accents
lowerCamelCase_ = tokenize_chinese_chars
lowerCamelCase_ = normalizer_class(**lowercase )
lowerCamelCase_ = do_lower_case
def SCREAMING_SNAKE_CASE_( self , lowercase , lowercase=None ) -> int:
lowerCamelCase_ = [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 , lowercase , lowercase = None ) -> List[int]:
lowerCamelCase_ = [self.sep_token_id]
lowerCamelCase_ = [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 , lowercase , lowercase = None ) -> Tuple[str]:
lowerCamelCase_ = self._tokenizer.model.save(lowercase , name=lowercase )
return tuple(lowercase )
| 19 |
"""simple docstring"""
import gc
import unittest
from diffusers import FlaxControlNetModel, FlaxStableDiffusionControlNetPipeline
from diffusers.utils import is_flax_available, load_image, slow
from diffusers.utils.testing_utils import require_flax
if is_flax_available():
import jax
import jax.numpy as jnp
from flax.jax_utils import replicate
from flax.training.common_utils import shard
@slow
@require_flax
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: List[Any] ) -> Any:
# clean up the VRAM after each test
super().tearDown()
gc.collect()
def lowerCAmelCase_ ( self: Tuple ) -> Any:
snake_case_, snake_case_ :List[str] = FlaxControlNetModel.from_pretrained(
"""lllyasviel/sd-controlnet-canny""" , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_, snake_case_ :Union[str, Any] = FlaxStableDiffusionControlNetPipeline.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" , controlnet=snake_case , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_ :Union[str, Any] = controlnet_params
snake_case_ :Union[str, Any] = """bird"""
snake_case_ :List[Any] = jax.device_count()
snake_case_ :List[Any] = pipe.prepare_text_inputs([prompts] * num_samples )
snake_case_ :List[str] = load_image(
"""https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/sd_controlnet/bird_canny.png""" )
snake_case_ :List[str] = pipe.prepare_image_inputs([canny_image] * num_samples )
snake_case_ :Any = jax.random.PRNGKey(0 )
snake_case_ :List[str] = jax.random.split(snake_case , jax.device_count() )
snake_case_ :List[Any] = replicate(snake_case )
snake_case_ :List[str] = shard(snake_case )
snake_case_ :str = shard(snake_case )
snake_case_ :Dict = pipe(
prompt_ids=snake_case , image=snake_case , params=snake_case , prng_seed=snake_case , num_inference_steps=50 , jit=snake_case , ).images
assert images.shape == (jax.device_count(), 1, 768, 512, 3)
snake_case_ :str = images.reshape((images.shape[0] * images.shape[1],) + images.shape[-3:] )
snake_case_ :Union[str, Any] = images[0, 253:256, 253:256, -1]
snake_case_ :str = jnp.asarray(jax.device_get(image_slice.flatten() ) )
snake_case_ :Dict = jnp.array(
[0.1_6_7_9_6_9, 0.1_1_6_6_9_9, 0.0_8_1_5_4_3, 0.1_5_4_2_9_7, 0.1_3_2_8_1_2, 0.1_0_8_8_8_7, 0.1_6_9_9_2_2, 0.1_6_9_9_2_2, 0.2_0_5_0_7_8] )
print(f"""output_slice: {output_slice}""" )
assert jnp.abs(output_slice - expected_slice ).max() < 1E-2
def lowerCAmelCase_ ( self: int ) -> Dict:
snake_case_, snake_case_ :List[Any] = FlaxControlNetModel.from_pretrained(
"""lllyasviel/sd-controlnet-openpose""" , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_, snake_case_ :int = FlaxStableDiffusionControlNetPipeline.from_pretrained(
"""runwayml/stable-diffusion-v1-5""" , controlnet=snake_case , from_pt=snake_case , dtype=jnp.bfloataa )
snake_case_ :str = controlnet_params
snake_case_ :Optional[int] = """Chef in the kitchen"""
snake_case_ :Union[str, Any] = jax.device_count()
snake_case_ :Any = pipe.prepare_text_inputs([prompts] * num_samples )
snake_case_ :str = load_image(
"""https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/sd_controlnet/pose.png""" )
snake_case_ :Optional[Any] = pipe.prepare_image_inputs([pose_image] * num_samples )
snake_case_ :str = jax.random.PRNGKey(0 )
snake_case_ :str = jax.random.split(snake_case , jax.device_count() )
snake_case_ :Tuple = replicate(snake_case )
snake_case_ :str = shard(snake_case )
snake_case_ :int = shard(snake_case )
snake_case_ :List[str] = pipe(
prompt_ids=snake_case , image=snake_case , params=snake_case , prng_seed=snake_case , num_inference_steps=50 , jit=snake_case , ).images
assert images.shape == (jax.device_count(), 1, 768, 512, 3)
snake_case_ :str = images.reshape((images.shape[0] * images.shape[1],) + images.shape[-3:] )
snake_case_ :int = images[0, 253:256, 253:256, -1]
snake_case_ :Dict = jnp.asarray(jax.device_get(image_slice.flatten() ) )
snake_case_ :Optional[int] = jnp.array(
[[0.2_7_1_4_8_4, 0.2_6_1_7_1_9, 0.2_7_5_3_9_1, 0.2_7_7_3_4_4, 0.2_7_9_2_9_7, 0.2_9_1_0_1_6, 0.2_9_4_9_2_2, 0.3_0_2_7_3_4, 0.3_0_2_7_3_4]] )
print(f"""output_slice: {output_slice}""" )
assert jnp.abs(output_slice - expected_slice ).max() < 1E-2
| 66 | 0 |
import json
import os
from typing import Optional, Tuple
from ...tokenization_utils import PreTrainedTokenizer
from ...utils import logging
lowercase : int = logging.get_logger(__name__)
lowercase : Optional[Any] = {"""vocab_file""": """vocab.json"""}
lowercase : Optional[Any] = {
"""vocab_file""": {
"""mgp-str""": """https://huggingface.co/alibaba-damo/mgp-str-base/blob/main/vocab.json""",
}
}
lowercase : int = {"""mgp-str""": 27}
class __snake_case ( lowerCAmelCase ):
_a : List[Any]= VOCAB_FILES_NAMES
_a : Optional[int]= PRETRAINED_VOCAB_FILES_MAP
_a : str= PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
def __init__( self ,snake_case ,snake_case="[GO]" ,snake_case="[GO]" ,snake_case="[s]" ,snake_case="[GO]" ,**snake_case ):
'''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:
lowercase : Dict = json.load(snake_case )
lowercase : Dict = {v: k for k, v in self.vocab.items()}
@property
def _SCREAMING_SNAKE_CASE ( self ):
'''simple docstring'''
return len(self.vocab )
def _SCREAMING_SNAKE_CASE ( self ):
'''simple docstring'''
return dict(self.vocab ,**self.added_tokens_encoder )
def _SCREAMING_SNAKE_CASE ( self ,snake_case ):
'''simple docstring'''
lowercase : Dict = []
for s in text:
char_tokens.extend(snake_case )
return char_tokens
def _SCREAMING_SNAKE_CASE ( self ,snake_case ):
'''simple docstring'''
return self.vocab.get(snake_case ,self.vocab.get(self.unk_token ) )
def _SCREAMING_SNAKE_CASE ( self ,snake_case ):
'''simple docstring'''
return self.decoder.get(snake_case )
def _SCREAMING_SNAKE_CASE ( self ,snake_case ,snake_case = None ):
'''simple docstring'''
if not os.path.isdir(snake_case ):
logger.error("""Vocabulary path ({}) should be a directory""".format(snake_case ) )
return
lowercase : Dict = os.path.join(
snake_case ,(filename_prefix + """-""" if filename_prefix else """""") + VOCAB_FILES_NAMES["""vocab_file"""] )
with open(snake_case ,"""w""" ,encoding="""utf-8""" ) as f:
f.write(json.dumps(self.vocab ,indent=2 ,sort_keys=snake_case ,ensure_ascii=snake_case ) + """\n""" )
return (vocab_file,)
| 20 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_tf_available,
is_tokenizers_available,
is_torch_available,
)
__a = {
"configuration_mobilebert": [
"MOBILEBERT_PRETRAINED_CONFIG_ARCHIVE_MAP",
"MobileBertConfig",
"MobileBertOnnxConfig",
],
"tokenization_mobilebert": ["MobileBertTokenizer"],
}
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["MobileBertTokenizerFast"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST",
"MobileBertForMaskedLM",
"MobileBertForMultipleChoice",
"MobileBertForNextSentencePrediction",
"MobileBertForPreTraining",
"MobileBertForQuestionAnswering",
"MobileBertForSequenceClassification",
"MobileBertForTokenClassification",
"MobileBertLayer",
"MobileBertModel",
"MobileBertPreTrainedModel",
"load_tf_weights_in_mobilebert",
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"TF_MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST",
"TFMobileBertForMaskedLM",
"TFMobileBertForMultipleChoice",
"TFMobileBertForNextSentencePrediction",
"TFMobileBertForPreTraining",
"TFMobileBertForQuestionAnswering",
"TFMobileBertForSequenceClassification",
"TFMobileBertForTokenClassification",
"TFMobileBertMainLayer",
"TFMobileBertModel",
"TFMobileBertPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_mobilebert import (
MOBILEBERT_PRETRAINED_CONFIG_ARCHIVE_MAP,
MobileBertConfig,
MobileBertOnnxConfig,
)
from .tokenization_mobilebert import MobileBertTokenizer
try:
if not is_tokenizers_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .tokenization_mobilebert_fast import MobileBertTokenizerFast
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_mobilebert import (
MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST,
MobileBertForMaskedLM,
MobileBertForMultipleChoice,
MobileBertForNextSentencePrediction,
MobileBertForPreTraining,
MobileBertForQuestionAnswering,
MobileBertForSequenceClassification,
MobileBertForTokenClassification,
MobileBertLayer,
MobileBertModel,
MobileBertPreTrainedModel,
load_tf_weights_in_mobilebert,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_mobilebert import (
TF_MOBILEBERT_PRETRAINED_MODEL_ARCHIVE_LIST,
TFMobileBertForMaskedLM,
TFMobileBertForMultipleChoice,
TFMobileBertForNextSentencePrediction,
TFMobileBertForPreTraining,
TFMobileBertForQuestionAnswering,
TFMobileBertForSequenceClassification,
TFMobileBertForTokenClassification,
TFMobileBertMainLayer,
TFMobileBertModel,
TFMobileBertPreTrainedModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 | 0 |
import os
from typing import List, Optional, Union
from ...image_processing_utils import BatchFeature
from ...image_utils import ImageInput
from ...processing_utils import ProcessorMixin
from ...tokenization_utils_base import PaddingStrategy, PreTokenizedInput, TextInput, TruncationStrategy
from ...utils import TensorType
from ..auto import AutoTokenizer
class _lowerCamelCase( _a ):
lowercase_ : Optional[int] = ["""image_processor""", """tokenizer"""]
lowercase_ : Dict = """BlipImageProcessor"""
lowercase_ : List[str] = """AutoTokenizer"""
def __init__( self, lowerCamelCase, lowerCamelCase, lowerCamelCase) -> List[Any]:
"""simple docstring"""
super().__init__(lowerCamelCase, lowerCamelCase)
# add QFormer tokenizer
_lowercase : int = qformer_tokenizer
def __call__( self, lowerCamelCase = None, lowerCamelCase = None, lowerCamelCase = True, lowerCamelCase = False, lowerCamelCase = None, lowerCamelCase = None, lowerCamelCase = 0, lowerCamelCase = None, lowerCamelCase = None, lowerCamelCase = False, lowerCamelCase = False, lowerCamelCase = False, lowerCamelCase = False, lowerCamelCase = False, lowerCamelCase = True, lowerCamelCase = None, **lowerCamelCase, ) -> BatchFeature:
"""simple docstring"""
if images is None and text is None:
raise ValueError('You have to specify at least images or text.')
_lowercase : int = BatchFeature()
if text is not None:
_lowercase : Union[str, Any] = self.tokenizer(
text=lowerCamelCase, add_special_tokens=lowerCamelCase, padding=lowerCamelCase, truncation=lowerCamelCase, max_length=lowerCamelCase, stride=lowerCamelCase, pad_to_multiple_of=lowerCamelCase, return_attention_mask=lowerCamelCase, return_overflowing_tokens=lowerCamelCase, return_special_tokens_mask=lowerCamelCase, return_offsets_mapping=lowerCamelCase, return_token_type_ids=lowerCamelCase, return_length=lowerCamelCase, verbose=lowerCamelCase, return_tensors=lowerCamelCase, **lowerCamelCase, )
encoding.update(lowerCamelCase)
_lowercase : str = self.qformer_tokenizer(
text=lowerCamelCase, add_special_tokens=lowerCamelCase, padding=lowerCamelCase, truncation=lowerCamelCase, max_length=lowerCamelCase, stride=lowerCamelCase, pad_to_multiple_of=lowerCamelCase, return_attention_mask=lowerCamelCase, return_overflowing_tokens=lowerCamelCase, return_special_tokens_mask=lowerCamelCase, return_offsets_mapping=lowerCamelCase, return_token_type_ids=lowerCamelCase, return_length=lowerCamelCase, verbose=lowerCamelCase, return_tensors=lowerCamelCase, **lowerCamelCase, )
_lowercase : Tuple = qformer_text_encoding.pop('input_ids')
_lowercase : Tuple = qformer_text_encoding.pop('attention_mask')
if images is not None:
_lowercase : Any = self.image_processor(lowerCamelCase, return_tensors=lowerCamelCase)
encoding.update(lowerCamelCase)
return encoding
def UpperCamelCase ( self, *lowerCamelCase, **lowerCamelCase) -> int:
"""simple docstring"""
return self.tokenizer.batch_decode(*lowerCamelCase, **lowerCamelCase)
def UpperCamelCase ( self, *lowerCamelCase, **lowerCamelCase) -> int:
"""simple docstring"""
return self.tokenizer.decode(*lowerCamelCase, **lowerCamelCase)
@property
# Copied from transformers.models.blip.processing_blip.BlipProcessor.model_input_names
def UpperCamelCase ( self) -> Union[str, Any]:
"""simple docstring"""
_lowercase : Optional[int] = self.tokenizer.model_input_names
_lowercase : Union[str, Any] = self.image_processor.model_input_names
return list(dict.fromkeys(tokenizer_input_names + image_processor_input_names))
def UpperCamelCase ( self, lowerCamelCase, **lowerCamelCase) -> Optional[Any]:
"""simple docstring"""
if os.path.isfile(lowerCamelCase):
raise ValueError(F'''Provided path ({save_directory}) should be a directory, not a file''')
os.makedirs(lowerCamelCase, exist_ok=lowerCamelCase)
_lowercase : Tuple = os.path.join(lowerCamelCase, 'qformer_tokenizer')
self.qformer_tokenizer.save_pretrained(lowerCamelCase)
return super().save_pretrained(lowerCamelCase, **lowerCamelCase)
@classmethod
def UpperCamelCase ( cls, lowerCamelCase, **lowerCamelCase) -> str:
"""simple docstring"""
_lowercase : Optional[Any] = AutoTokenizer.from_pretrained(lowerCamelCase, subfolder='qformer_tokenizer')
_lowercase : Any = cls._get_arguments_from_pretrained(lowerCamelCase, **lowerCamelCase)
args.append(lowerCamelCase)
return cls(*lowerCamelCase)
| 21 |
"""simple docstring"""
import argparse
import json
import os
from collections import OrderedDict
import numpy as np
import tensorflow as tf
import torch
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Union[str, Any] = os.path.join(args.tf_model_dir, """parameters.json""" )
snake_case_ :Any = json.loads(open(_lowercase ).read() )
if not params:
raise ValueError(
f"""It seems that the json file at {parameter_file} is empty. Make sure you have a correct json file.""" )
if not args.output.endswith(""".pt""" ):
snake_case_ :Optional[int] = args.output + """.pt"""
snake_case_ :List[str] = OrderedDict()
with tf.device("""/CPU:0""" ):
snake_case_ :Dict = tf.train.load_checkpoint(args.tf_model_dir )
snake_case_ :str = reader.get_variable_to_shape_map()
for key_name in shapes.keys():
snake_case_ :List[Any] = reader.get_tensor(_lowercase ).astype(np.floataa )
if key_name.endswith("""/adam_m""" ) or key_name.endswith("""/adam_v""" ):
continue
if key_name.startswith("""pasts/""" ):
if key_name.startswith("""pasts/mlp""" ):
snake_case_ :Any = int(key_name[9] )
elif key_name.startswith("""pasts/out""" ):
snake_case_ :Optional[int] = 8
snake_case_ :List[str] = """model.sqout.%d.weight""" % (player * 2) # enter to nn.Sequencial with Tanh, so 2 at a time
snake_case_ :Optional[Any] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :List[str] = torch.tensor(_lowercase )
elif key_name.startswith("""model/moe""" ):
snake_case_ :Tuple = int(key_name[9:].split("""/""" )[0] )
if key_name.endswith("""/switch_gating/kernel""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.feed_forward.mlp.router.classifier.weight""" % player
snake_case_ :Optional[Any] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/softmlp/kernel""" ):
snake_case_ :List[Any] = """model.blocks.%d.feed_forward.soft_bypass_mlp.weight""" % player
snake_case_ :Optional[int] = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/wo/kernel""" ) or key_name.endswith("""/wi/kernel""" ):
snake_case_ :Dict = key_name[-9:-7]
for i in range(16 ):
snake_case_ :str = """model.blocks.%d.feed_forward.mlp.experts.expert_%d.%s.weight""" % (player, i, nlayer)
snake_case_ :Tuple = (
vnp[i].transpose([1, 0] ).copy()
) # In Mesh-Tensorflow, it is one array, so it is divided
snake_case_ :Optional[int] = torch.tensor(_lowercase )
elif key_name.startswith("""model/mlp""" ):
snake_case_ :Optional[int] = int(key_name[9:].split("""/""" )[0] )
if key_name.endswith("""/p1/kernel""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.feed_forward.mlp.wi.weight""" % player
snake_case_ :Dict = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/p1/bias""" ):
snake_case_ :List[Any] = """model.blocks.%d.feed_forward.mlp.wi.bias""" % player
snake_case_ :str = vnp.copy() # same because it is one dimensional
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/p2/kernel""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.feed_forward.mlp.wo.weight""" % player
snake_case_ :Dict = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Tuple = torch.tensor(_lowercase )
elif key_name.endswith("""/p2/bias""" ):
snake_case_ :Dict = """model.blocks.%d.feed_forward.mlp.wo.bias""" % player
snake_case_ :Any = vnp.copy() # same because it is one dimensional
snake_case_ :Optional[int] = torch.tensor(_lowercase )
elif key_name.startswith("""model/ln""" ):
snake_case_ :Union[str, Any] = int(key_name[8:].split("""/""" )[0] )
if key_name.endswith("""/b""" ):
snake_case_ :str = """model.blocks.%d.feed_forward.norm.bias""" % player
snake_case_ :Dict = vnp.copy() # same because it is one dimensional
snake_case_ :int = torch.tensor(_lowercase )
elif key_name.endswith("""/g""" ):
snake_case_ :Dict = """model.blocks.%d.feed_forward.norm.weight""" % player
snake_case_ :Dict = vnp.copy() # same because it is one dimensional
snake_case_ :Tuple = torch.tensor(_lowercase )
elif key_name.startswith("""model/att""" ):
snake_case_ :List[str] = int(key_name[9:].split("""/""" )[0] )
if key_name.endswith("""/qkv/kernel""" ):
snake_case_ :Optional[int] = vnp.copy() # Compute same dimension as Mesh-tensorflow using einsum
snake_case_ :Dict = state[:, 0, :, :]
snake_case_ :int = state[:, 1, :, :]
snake_case_ :List[str] = state[:, 2, :, :]
snake_case_ :str = (
state_q.reshape([state_q.shape[0], state_q.shape[1] * state_q.shape[2]] )
.transpose([1, 0] )
.copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Any = (
state_k.reshape([state_k.shape[0], state_k.shape[1] * state_k.shape[2]] )
.transpose([1, 0] )
.copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Optional[int] = (
state_v.reshape([state_v.shape[0], state_v.shape[1] * state_v.shape[2]] )
.transpose([1, 0] )
.copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :int = """model.blocks.%d.self_attn.self_attn.q_proj.weight""" % player
snake_case_ :int = torch.tensor(_lowercase )
snake_case_ :Optional[Any] = """model.blocks.%d.self_attn.self_attn.k_proj.weight""" % player
snake_case_ :Dict = torch.tensor(_lowercase )
snake_case_ :Dict = """model.blocks.%d.self_attn.self_attn.v_proj.weight""" % player
snake_case_ :Optional[Any] = torch.tensor(_lowercase )
elif key_name.endswith("""/o/kernel""" ):
snake_case_ :str = """model.blocks.%d.self_attn.self_attn.out_proj.weight""" % player
snake_case_ :str = (
vnp.reshape([vnp.shape[0] * vnp.shape[1], vnp.shape[2]] ).transpose([1, 0] ).copy()
) # Mesh-Tensorflow is a diagonal matrix
snake_case_ :Any = torch.tensor(_lowercase )
elif key_name.startswith("""model/an""" ):
snake_case_ :Optional[int] = int(key_name[8:].split("""/""" )[0] )
if key_name.endswith("""/b""" ):
snake_case_ :Any = """model.blocks.%d.self_attn.norm.bias""" % player
snake_case_ :Optional[int] = vnp.copy() # same because it is one dimensional
snake_case_ :Tuple = torch.tensor(_lowercase )
elif key_name.endswith("""/g""" ):
snake_case_ :Union[str, Any] = """model.blocks.%d.self_attn.norm.weight""" % player
snake_case_ :Dict = vnp.copy() # same because it is one dimensional
snake_case_ :Optional[int] = torch.tensor(_lowercase )
elif (
key_name.startswith("""model/wte""" )
or key_name.startswith("""model/wpe""" )
or key_name.startswith("""model/ete""" )
):
snake_case_ :List[Any] = {"""wte""": """embed_tokens""", """wpe""": """position_embeddings""", """ete""": """extra_position_embeddings"""}[
key_name[-3:]
]
snake_case_ :Optional[Any] = """model.%s.weight""" % nlayer
snake_case_ :Any = vnp.copy() # same in embedded
snake_case_ :List[Any] = torch.tensor(_lowercase )
if key_name.startswith("""model/wte""" ):
snake_case_ :Tuple = """lm_head.weight"""
snake_case_ :List[str] = vnp.copy() # same in embedded
snake_case_ :List[Any] = torch.tensor(_lowercase )
elif key_name.startswith("""model/wob""" ):
snake_case_ :str = """final_logits_bias"""
snake_case_ :Any = vnp.copy() # same in embedded
snake_case_ :List[Any] = state.reshape((1, -1) )
snake_case_ :Union[str, Any] = torch.tensor(_lowercase )
elif key_name == "model/dense/kernel":
snake_case_ :str = """model.last_project.weight"""
snake_case_ :Dict = vnp.transpose([1, 0] ).copy() # Mesh-Tensorflow is a diagonal matrix
snake_case_ :int = torch.tensor(_lowercase )
elif key_name == "model/dense_1/bias":
snake_case_ :Optional[int] = """model.last_project.bias"""
snake_case_ :Tuple = vnp.copy() # same because it is one dimensional
snake_case_ :Any = torch.tensor(_lowercase )
torch.save(_lowercase, args.output )
if __name__ == "__main__":
__a = argparse.ArgumentParser(
description="model converter.", formatter_class=argparse.ArgumentDefaultsHelpFormatter
)
parser.add_argument("--tf_model_dir", metavar="PATH", type=str, required=True, help="import model")
parser.add_argument("--output", metavar="PATH", type=str, required=True, help="output model")
__a = parser.parse_args()
convert_tf_gptsan_to_pt(args)
| 66 | 0 |
'''simple docstring'''
from __future__ import annotations
import unittest
from transformers import is_tf_available, is_torch_available
from transformers.testing_utils import DUMMY_UNKNOWN_IDENTIFIER, SMALL_MODEL_IDENTIFIER, is_pt_tf_cross_test, slow
if is_tf_available():
from transformers import (
AutoConfig,
BertConfig,
GPTaConfig,
TaConfig,
TFAutoModel,
TFAutoModelForCausalLM,
TFAutoModelForMaskedLM,
TFAutoModelForPreTraining,
TFAutoModelForQuestionAnswering,
TFAutoModelForSeqaSeqLM,
TFAutoModelForSequenceClassification,
TFAutoModelWithLMHead,
TFBertForMaskedLM,
TFBertForPreTraining,
TFBertForQuestionAnswering,
TFBertForSequenceClassification,
TFBertModel,
TFGPTaLMHeadModel,
TFRobertaForMaskedLM,
TFTaForConditionalGeneration,
)
from transformers.models.bert.modeling_tf_bert import TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST
from transformers.models.gpta.modeling_tf_gpta import TF_GPT2_PRETRAINED_MODEL_ARCHIVE_LIST
from transformers.models.ta.modeling_tf_ta import TF_T5_PRETRAINED_MODEL_ARCHIVE_LIST
if is_torch_available():
from transformers import (
AutoModel,
AutoModelForCausalLM,
AutoModelForMaskedLM,
AutoModelForPreTraining,
AutoModelForQuestionAnswering,
AutoModelForSeqaSeqLM,
AutoModelForSequenceClassification,
AutoModelWithLMHead,
BertForMaskedLM,
BertForPreTraining,
BertForQuestionAnswering,
BertForSequenceClassification,
BertModel,
GPTaLMHeadModel,
RobertaForMaskedLM,
TaForConditionalGeneration,
)
@is_pt_tf_cross_test
class A_ ( unittest.TestCase ):
@slow
def lowercase ( self : List[Any] ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
_UpperCAmelCase = AutoConfig.from_pretrained(snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = TFAutoModel.from_pretrained(snake_case_ , from_pt=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = AutoModel.from_pretrained(snake_case_ , from_tf=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
@slow
def lowercase ( self : str ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
_UpperCAmelCase = AutoConfig.from_pretrained(snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = TFAutoModelForPreTraining.from_pretrained(snake_case_ , from_pt=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = AutoModelForPreTraining.from_pretrained(snake_case_ , from_tf=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
@slow
def lowercase ( self : List[str] ):
for model_name in TF_GPT2_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
_UpperCAmelCase = AutoConfig.from_pretrained(snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = TFAutoModelForCausalLM.from_pretrained(snake_case_ , from_pt=snake_case_ )
_UpperCAmelCase , _UpperCAmelCase = TFAutoModelForCausalLM.from_pretrained(
snake_case_ , output_loading_info=snake_case_ , from_pt=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = AutoModelForCausalLM.from_pretrained(snake_case_ , from_tf=snake_case_ )
_UpperCAmelCase , _UpperCAmelCase = AutoModelForCausalLM.from_pretrained(
snake_case_ , output_loading_info=snake_case_ , from_tf=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
@slow
def lowercase ( self : Union[str, Any] ):
for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
_UpperCAmelCase = AutoConfig.from_pretrained(snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = TFAutoModelWithLMHead.from_pretrained(snake_case_ , from_pt=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = AutoModelWithLMHead.from_pretrained(snake_case_ , from_tf=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
@slow
def lowercase ( self : Any ):
for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
_UpperCAmelCase = AutoConfig.from_pretrained(snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = TFAutoModelForMaskedLM.from_pretrained(snake_case_ , from_pt=snake_case_ )
_UpperCAmelCase , _UpperCAmelCase = TFAutoModelForMaskedLM.from_pretrained(
snake_case_ , output_loading_info=snake_case_ , from_pt=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = AutoModelForMaskedLM.from_pretrained(snake_case_ , from_tf=snake_case_ )
_UpperCAmelCase , _UpperCAmelCase = AutoModelForMaskedLM.from_pretrained(
snake_case_ , output_loading_info=snake_case_ , from_tf=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
@slow
def lowercase ( self : int ):
for model_name in TF_T5_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
_UpperCAmelCase = AutoConfig.from_pretrained(snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = TFAutoModelForSeqaSeqLM.from_pretrained(snake_case_ , from_pt=snake_case_ )
_UpperCAmelCase , _UpperCAmelCase = TFAutoModelForSeqaSeqLM.from_pretrained(
snake_case_ , output_loading_info=snake_case_ , from_pt=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = AutoModelForSeqaSeqLM.from_pretrained(snake_case_ , from_tf=snake_case_ )
_UpperCAmelCase , _UpperCAmelCase = AutoModelForSeqaSeqLM.from_pretrained(
snake_case_ , output_loading_info=snake_case_ , from_tf=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
@slow
def lowercase ( self : Tuple ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
_UpperCAmelCase = AutoConfig.from_pretrained(snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = TFAutoModelForSequenceClassification.from_pretrained(snake_case_ , from_pt=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = AutoModelForSequenceClassification.from_pretrained(snake_case_ , from_tf=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
@slow
def lowercase ( self : Tuple ):
# for model_name in TF_BERT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
for model_name in ["bert-base-uncased"]:
_UpperCAmelCase = AutoConfig.from_pretrained(snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = TFAutoModelForQuestionAnswering.from_pretrained(snake_case_ , from_pt=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
_UpperCAmelCase = AutoModelForQuestionAnswering.from_pretrained(snake_case_ , from_tf=snake_case_ )
self.assertIsNotNone(snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
def lowercase ( self : Union[str, Any] ):
_UpperCAmelCase = TFAutoModelWithLMHead.from_pretrained(snake_case_ , from_pt=snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=snake_case_ ) , 1_4_4_1_0 )
_UpperCAmelCase = AutoModelWithLMHead.from_pretrained(snake_case_ , from_tf=snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=snake_case_ ) , 1_4_4_1_0 )
def lowercase ( self : List[Any] ):
_UpperCAmelCase = TFAutoModelWithLMHead.from_pretrained(snake_case_ , from_pt=snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=snake_case_ ) , 1_4_4_1_0 )
_UpperCAmelCase = AutoModelWithLMHead.from_pretrained(snake_case_ , from_tf=snake_case_ )
self.assertIsInstance(snake_case_ , snake_case_ )
self.assertEqual(model.num_parameters() , 1_4_4_1_0 )
self.assertEqual(model.num_parameters(only_trainable=snake_case_ ) , 1_4_4_1_0 )
| 22 |
"""simple docstring"""
import numpy as np
import pandas as pd
from sklearn.preprocessing import MinMaxScaler
from tensorflow.keras.layers import LSTM, Dense
from tensorflow.keras.models import Sequential
if __name__ == "__main__":
__a = pd.read_csv("sample_data.csv", header=None)
__a = df.shape[:1][0]
# If you're using some other dataset input the target column
__a = df.iloc[:, 1:2]
__a = actual_data.values.reshape(len_data, 1)
__a = MinMaxScaler().fit_transform(actual_data)
__a = 10
__a = 5
__a = 20
__a = len_data - periods * look_back
__a = actual_data[:division]
__a = actual_data[division - look_back :]
__a , __a = [], []
__a , __a = [], []
for i in range(0, len(train_data) - forward_days - look_back + 1):
train_x.append(train_data[i : i + look_back])
train_y.append(train_data[i + look_back : i + look_back + forward_days])
for i in range(0, len(test_data) - forward_days - look_back + 1):
test_x.append(test_data[i : i + look_back])
test_y.append(test_data[i + look_back : i + look_back + forward_days])
__a = np.array(train_x)
__a = np.array(test_x)
__a = np.array([list(i.ravel()) for i in train_y])
__a = np.array([list(i.ravel()) for i in test_y])
__a = Sequential()
model.add(LSTM(1_28, input_shape=(look_back, 1), return_sequences=True))
model.add(LSTM(64, input_shape=(1_28, 1)))
model.add(Dense(forward_days))
model.compile(loss="mean_squared_error", optimizer="adam")
__a = model.fit(
x_train, y_train, epochs=1_50, verbose=1, shuffle=True, batch_size=4
)
__a = model.predict(x_test)
| 66 | 0 |
'''simple docstring'''
import warnings
from ...utils import logging
from .image_processing_poolformer import PoolFormerImageProcessor
UpperCamelCase__: Union[str, Any] = logging.get_logger(__name__)
class SCREAMING_SNAKE_CASE( A__ ):
"""simple docstring"""
def __init__( self : str , *__snake_case : str , **__snake_case : int ) -> None:
warnings.warn(
'''The class PoolFormerFeatureExtractor is deprecated and will be removed in version 5 of Transformers.'''
''' Please use PoolFormerImageProcessor instead.''' , __snake_case , )
super().__init__(*__snake_case , **__snake_case )
| 23 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_tokenizers_available, is_torch_available
__a = {
"configuration_altclip": [
"ALTCLIP_PRETRAINED_CONFIG_ARCHIVE_MAP",
"AltCLIPConfig",
"AltCLIPTextConfig",
"AltCLIPVisionConfig",
],
"processing_altclip": ["AltCLIPProcessor"],
}
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"ALTCLIP_PRETRAINED_MODEL_ARCHIVE_LIST",
"AltCLIPPreTrainedModel",
"AltCLIPModel",
"AltCLIPTextModel",
"AltCLIPVisionModel",
]
if TYPE_CHECKING:
from .configuration_altclip import (
ALTCLIP_PRETRAINED_CONFIG_ARCHIVE_MAP,
AltCLIPConfig,
AltCLIPTextConfig,
AltCLIPVisionConfig,
)
from .processing_altclip import AltCLIPProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_altclip import (
ALTCLIP_PRETRAINED_MODEL_ARCHIVE_LIST,
AltCLIPModel,
AltCLIPPreTrainedModel,
AltCLIPTextModel,
AltCLIPVisionModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)
| 66 | 0 |
import os
import unittest
from transformers.models.bartpho.tokenization_bartpho import VOCAB_FILES_NAMES, BartphoTokenizer
from transformers.testing_utils import get_tests_dir
from ...test_tokenization_common import TokenizerTesterMixin
snake_case_ = get_tests_dir('fixtures/test_sentencepiece_bpe.model')
class SCREAMING_SNAKE_CASE__ ( _UpperCAmelCase , unittest.TestCase ):
A_ : List[Any] = BartphoTokenizer
A_ : List[str] = False
A_ : Optional[Any] = True
def a (self : Tuple ):
"""simple docstring"""
super().setUp()
__snake_case = ['''▁This''', '''▁is''', '''▁a''', '''▁t''', '''est''']
__snake_case = dict(zip(a__ , range(len(a__ ) ) ) )
__snake_case = {'''unk_token''': '''<unk>'''}
__snake_case = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['''monolingual_vocab_file'''] )
with open(self.monolingual_vocab_file , '''w''' , encoding='''utf-8''' ) as fp:
for token in vocab_tokens:
fp.write(f"""{token} {vocab_tokens[token]}\n""" )
__snake_case = BartphoTokenizer(a__ , self.monolingual_vocab_file , **self.special_tokens_map )
tokenizer.save_pretrained(self.tmpdirname )
def a (self : str , **a__ : str ):
"""simple docstring"""
kwargs.update(self.special_tokens_map )
return BartphoTokenizer.from_pretrained(self.tmpdirname , **a__ )
def a (self : str , a__ : Any ):
"""simple docstring"""
__snake_case = '''This is a là test'''
__snake_case = '''This is a<unk><unk> test'''
return input_text, output_text
def a (self : Dict ):
"""simple docstring"""
__snake_case = BartphoTokenizer(a__ , self.monolingual_vocab_file , **self.special_tokens_map )
__snake_case = '''This is a là test'''
__snake_case = '''▁This ▁is ▁a ▁l à ▁t est'''.split()
__snake_case = tokenizer.tokenize(a__ )
self.assertListEqual(a__ , a__ )
__snake_case = tokens + [tokenizer.unk_token]
__snake_case = [4, 5, 6, 3, 3, 7, 8, 3]
self.assertListEqual(tokenizer.convert_tokens_to_ids(a__ ) , a__ )
| 24 |
"""simple docstring"""
import argparse
import gdown
import numpy as np
import torch
from huggingface_hub import hf_hub_download
from transformers import (
CLIPTokenizer,
CLIPTokenizerFast,
VideoMAEImageProcessor,
XCLIPConfig,
XCLIPModel,
XCLIPProcessor,
XCLIPTextConfig,
XCLIPVisionConfig,
)
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :int = XCLIPTextConfig()
# derive patch size from model name
snake_case_ :Union[str, Any] = model_name.find("""patch""" )
snake_case_ :List[str] = int(model_name[start_idx + len("""patch""" ) : start_idx + len("""patch""" ) + 2] )
snake_case_ :Any = XCLIPVisionConfig(patch_size=_lowercase, num_frames=_lowercase )
if "large" in model_name:
snake_case_ :Optional[Any] = 768
snake_case_ :Union[str, Any] = 3072
snake_case_ :Any = 12
snake_case_ :Any = 1024
snake_case_ :str = 4096
snake_case_ :Union[str, Any] = 16
snake_case_ :Union[str, Any] = 24
snake_case_ :Tuple = 768
snake_case_ :Any = 3072
if model_name == "xclip-large-patch14-16-frames":
snake_case_ :Any = 336
snake_case_ :Any = XCLIPConfig.from_text_vision_configs(_lowercase, _lowercase )
if "large" in model_name:
snake_case_ :List[Any] = 768
return config
def A_ ( _lowercase ):
'''simple docstring'''
if name == "token_embedding.weight":
snake_case_ :Optional[Any] = name.replace("""token_embedding.weight""", """text_model.embeddings.token_embedding.weight""" )
if name == "positional_embedding":
snake_case_ :Tuple = name.replace("""positional_embedding""", """text_model.embeddings.position_embedding.weight""" )
if "ln_1" in name:
snake_case_ :Dict = name.replace("""ln_1""", """layer_norm1""" )
if "ln_2" in name:
snake_case_ :str = name.replace("""ln_2""", """layer_norm2""" )
if "c_fc" in name:
snake_case_ :str = name.replace("""c_fc""", """fc1""" )
if "c_proj" in name:
snake_case_ :int = name.replace("""c_proj""", """fc2""" )
if name.startswith("""transformer.resblocks""" ):
snake_case_ :Union[str, Any] = name.replace("""transformer.resblocks""", """text_model.encoder.layers""" )
if "attn.out_proj" in name and "message" not in name:
snake_case_ :Union[str, Any] = name.replace("""attn.out_proj""", """self_attn.out_proj""" )
if "ln_final" in name:
snake_case_ :Union[str, Any] = name.replace("""ln_final""", """text_model.final_layer_norm""" )
# visual encoder
if name == "visual.class_embedding":
snake_case_ :Any = name.replace("""visual.class_embedding""", """vision_model.embeddings.class_embedding""" )
if name == "visual.positional_embedding":
snake_case_ :Optional[int] = name.replace("""visual.positional_embedding""", """vision_model.embeddings.position_embedding.weight""" )
if name.startswith("""visual.transformer.resblocks""" ):
snake_case_ :Union[str, Any] = name.replace("""visual.transformer.resblocks""", """vision_model.encoder.layers""" )
if "visual.conv1" in name:
snake_case_ :int = name.replace("""visual.conv1""", """vision_model.embeddings.patch_embedding""" )
if "visual.ln_pre" in name:
snake_case_ :Any = name.replace("""visual.ln_pre""", """vision_model.pre_layernorm""" )
if "visual.ln_post" in name:
snake_case_ :str = name.replace("""visual.ln_post""", """vision_model.post_layernorm""" )
if "visual.proj" in name:
snake_case_ :Union[str, Any] = name.replace("""visual.proj""", """visual_projection.weight""" )
if "text_projection" in name:
snake_case_ :Dict = name.replace("""text_projection""", """text_projection.weight""" )
# things on top
if "prompts_visual_proj" in name:
snake_case_ :List[str] = name.replace("""prompts_visual_proj""", """prompts_visual_projection""" )
if "prompts_visual_ln" in name:
snake_case_ :Dict = name.replace("""prompts_visual_ln""", """prompts_visual_layernorm""" )
# mit
if name == "mit.positional_embedding":
snake_case_ :str = name.replace("""positional""", """position""" )
if name.startswith("""mit.resblocks""" ):
snake_case_ :Dict = name.replace("""mit.resblocks""", """mit.encoder.layers""" )
# prompts generator
if name.startswith("""prompts_generator.norm""" ):
snake_case_ :Union[str, Any] = name.replace("""prompts_generator.norm""", """prompts_generator.layernorm""" )
return name
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
for key in orig_state_dict.copy().keys():
snake_case_ :Dict = orig_state_dict.pop(_lowercase )
if "attn.in_proj" in key:
snake_case_ :Optional[Any] = key.split(""".""" )
if key.startswith("""visual""" ):
snake_case_ :Any = key_split[3]
snake_case_ :Optional[Any] = config.vision_config.hidden_size
if "message_attn" in key:
if "weight" in key:
snake_case_ :str = val[
:dim, :
]
snake_case_ :Optional[int] = val[
dim : dim * 2, :
]
snake_case_ :Union[str, Any] = val[
-dim:, :
]
else:
snake_case_ :Dict = val[
:dim
]
snake_case_ :Optional[int] = val[
dim : dim * 2
]
snake_case_ :Optional[int] = val[
-dim:
]
else:
if "weight" in key:
snake_case_ :Optional[Any] = val[
:dim, :
]
snake_case_ :List[str] = val[
dim : dim * 2, :
]
snake_case_ :Dict = val[
-dim:, :
]
else:
snake_case_ :Union[str, Any] = val[:dim]
snake_case_ :Union[str, Any] = val[
dim : dim * 2
]
snake_case_ :Union[str, Any] = val[-dim:]
elif key.startswith("""mit""" ):
snake_case_ :Tuple = key_split[2]
snake_case_ :Union[str, Any] = config.vision_config.mit_hidden_size
if "weight" in key:
snake_case_ :Optional[int] = val[:dim, :]
snake_case_ :Optional[int] = val[dim : dim * 2, :]
snake_case_ :str = val[-dim:, :]
else:
snake_case_ :str = val[:dim]
snake_case_ :Any = val[dim : dim * 2]
snake_case_ :int = val[-dim:]
else:
snake_case_ :Tuple = key_split[2]
snake_case_ :Any = config.text_config.hidden_size
if "weight" in key:
snake_case_ :Dict = val[:dim, :]
snake_case_ :Dict = val[
dim : dim * 2, :
]
snake_case_ :List[str] = val[-dim:, :]
else:
snake_case_ :Any = val[:dim]
snake_case_ :Tuple = val[
dim : dim * 2
]
snake_case_ :List[str] = val[-dim:]
else:
snake_case_ :Optional[int] = rename_key(_lowercase )
if new_key_name in ["visual_projection.weight", "text_projection.weight"]:
snake_case_ :Optional[Any] = val.T
snake_case_ :Tuple = val
return orig_state_dict
def A_ ( _lowercase ):
'''simple docstring'''
if num_frames == 8:
snake_case_ :str = """eating_spaghetti_8_frames.npy"""
elif num_frames == 16:
snake_case_ :int = """eating_spaghetti.npy"""
elif num_frames == 32:
snake_case_ :List[str] = """eating_spaghetti_32_frames.npy"""
snake_case_ :int = hf_hub_download(
repo_id="""hf-internal-testing/spaghetti-video""", filename=_lowercase, repo_type="""dataset""", )
snake_case_ :Union[str, Any] = np.load(_lowercase )
return list(_lowercase )
def A_ ( _lowercase, _lowercase=None, _lowercase=False ):
'''simple docstring'''
snake_case_ :List[Any] = {
# fully supervised kinetics-400 checkpoints
"""xclip-base-patch32""": """https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_32_8.pth""",
"""xclip-base-patch32-16-frames""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_32_16.pth"""
),
"""xclip-base-patch16""": """https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_16_8.pth""",
"""xclip-base-patch16-16-frames""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k400_16_16.pth"""
),
"""xclip-large-patch14""": """https://drive.google.com/u/0/uc?id=1NUOImq0o5DlQTST17iIP3vG7DgmHQuCx&export=download&confirm=t&uuid=b26caedc-88e2-473e-830a-9d158b653cdb""",
"""xclip-large-patch14-16-frames""": """https://drive.google.com/u/0/uc?id=1FOYgnJc097OJ4lGwtRCCydQyVPJEOH7d&export=download&confirm=t&uuid=538fa810-e671-4050-b385-9a623f89804f""",
# fully supervised kinetics-600 checkpoints
"""xclip-base-patch16-kinetics-600""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k600_16_8.pth"""
),
"""xclip-base-patch16-kinetics-600-16-frames""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/k600_16_16.pth"""
),
"""xclip-large-patch14-kinetics-600""": """https://drive.google.com/u/0/uc?id=1FV8C1INuM91sLAN4ImjzePLIlpMSihwV&export=download&confirm=t&uuid=141d4977-4a65-44ae-864f-4b0c19f838be""",
# few shot
"""xclip-base-patch16-hmdb-2-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_2.pth"""
),
"""xclip-base-patch16-hmdb-4-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_4.pth"""
),
"""xclip-base-patch16-hmdb-8-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_8.pth"""
),
"""xclip-base-patch16-hmdb-16-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_hmdb_16.pth"""
),
"""xclip-base-patch16-ucf-2-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_2.pth"""
),
"""xclip-base-patch16-ucf-4-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_4.pth"""
),
"""xclip-base-patch16-ucf-8-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_8.pth"""
),
"""xclip-base-patch16-ucf-16-shot""": (
"""https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/few_ucf_16.pth"""
),
# zero shot
"""xclip-base-patch16-zero-shot""": """https://github.com/nbl97/X-CLIP_Model_Zoo/releases/download/v1.0/zero.pth""",
}
snake_case_ :Optional[int] = model_to_url[model_name]
snake_case_ :int = 8
if "16-frames" in model_name:
snake_case_ :List[Any] = 16
elif "shot" in model_name:
snake_case_ :Dict = 32
snake_case_ :Optional[int] = get_xclip_config(_lowercase, _lowercase )
snake_case_ :Optional[Any] = XCLIPModel(_lowercase )
model.eval()
if "drive" in checkpoint_url:
snake_case_ :List[str] = """pytorch_model.bin"""
gdown.cached_download(_lowercase, _lowercase, quiet=_lowercase )
snake_case_ :List[Any] = torch.load(_lowercase, map_location="""cpu""" )["""model"""]
else:
snake_case_ :Tuple = torch.hub.load_state_dict_from_url(_lowercase )["""model"""]
snake_case_ :Union[str, Any] = convert_state_dict(_lowercase, _lowercase )
snake_case_ :str = XCLIPModel(_lowercase )
snake_case_, snake_case_ :Optional[int] = model.load_state_dict(_lowercase, strict=_lowercase )
assert missing_keys == ["text_model.embeddings.position_ids", "vision_model.embeddings.position_ids"]
model.eval()
snake_case_ :List[str] = 336 if model_name == """xclip-large-patch14-16-frames""" else 224
snake_case_ :List[Any] = VideoMAEImageProcessor(size=_lowercase )
snake_case_ :Any = CLIPTokenizer.from_pretrained("""openai/clip-vit-base-patch32""" )
snake_case_ :str = CLIPTokenizerFast.from_pretrained("""openai/clip-vit-base-patch32""" )
snake_case_ :Optional[Any] = XCLIPProcessor(image_processor=_lowercase, tokenizer=_lowercase )
snake_case_ :Optional[int] = prepare_video(_lowercase )
snake_case_ :Optional[Any] = processor(
text=["""playing sports""", """eating spaghetti""", """go shopping"""], videos=_lowercase, return_tensors="""pt""", padding=_lowercase )
print("""Shape of pixel values:""", inputs.pixel_values.shape )
with torch.no_grad():
snake_case_ :List[Any] = model(**_lowercase )
# Verify outputs
snake_case_ :List[Any] = outputs.logits_per_video
snake_case_ :Any = logits_per_video.softmax(dim=1 )
print("""Probs:""", _lowercase )
# kinetics-400
if model_name == "xclip-base-patch32":
snake_case_ :Union[str, Any] = torch.tensor([[0.0019, 0.9951, 0.0030]] )
elif model_name == "xclip-base-patch32-16-frames":
snake_case_ :str = torch.tensor([[7.09_99e-04, 9.98_83e-01, 4.55_80e-04]] )
elif model_name == "xclip-base-patch16":
snake_case_ :Tuple = torch.tensor([[0.0083, 0.9681, 0.0236]] )
elif model_name == "xclip-base-patch16-16-frames":
snake_case_ :Any = torch.tensor([[7.69_37e-04, 9.97_28e-01, 1.94_73e-03]] )
elif model_name == "xclip-large-patch14":
snake_case_ :str = torch.tensor([[0.0062, 0.9864, 0.0075]] )
elif model_name == "xclip-large-patch14-16-frames":
snake_case_ :Tuple = torch.tensor([[3.38_77e-04, 9.99_37e-01, 2.88_88e-04]] )
# kinetics-600
elif model_name == "xclip-base-patch16-kinetics-600":
snake_case_ :List[Any] = torch.tensor([[0.0555, 0.8914, 0.0531]] )
elif model_name == "xclip-base-patch16-kinetics-600-16-frames":
snake_case_ :Union[str, Any] = torch.tensor([[3.85_54e-04, 9.99_29e-01, 3.27_54e-04]] )
elif model_name == "xclip-large-patch14-kinetics-600":
snake_case_ :List[Any] = torch.tensor([[0.0036, 0.9920, 0.0045]] )
# few shot
elif model_name == "xclip-base-patch16-hmdb-2-shot":
snake_case_ :Dict = torch.tensor([[7.18_90e-06, 9.99_94e-01, 5.65_59e-05]] )
elif model_name == "xclip-base-patch16-hmdb-4-shot":
snake_case_ :Union[str, Any] = torch.tensor([[1.03_20e-05, 9.99_93e-01, 6.24_35e-05]] )
elif model_name == "xclip-base-patch16-hmdb-8-shot":
snake_case_ :str = torch.tensor([[4.13_77e-06, 9.99_90e-01, 9.83_86e-05]] )
elif model_name == "xclip-base-patch16-hmdb-16-shot":
snake_case_ :str = torch.tensor([[4.13_47e-05, 9.99_62e-01, 3.34_11e-04]] )
elif model_name == "xclip-base-patch16-ucf-2-shot":
snake_case_ :int = torch.tensor([[8.58_57e-05, 9.99_28e-01, 6.32_91e-04]] )
elif model_name == "xclip-base-patch16-ucf-4-shot":
snake_case_ :Optional[int] = torch.tensor([[8.58_57e-05, 9.99_28e-01, 6.32_91e-04]] )
elif model_name == "xclip-base-patch16-ucf-8-shot":
snake_case_ :Any = torch.tensor([[0.0027, 0.9904, 0.0070]] )
elif model_name == "xclip-base-patch16-ucf-16-shot":
snake_case_ :Tuple = torch.tensor([[9.82_19e-04, 9.95_93e-01, 3.08_63e-03]] )
# zero shot
elif model_name == "xclip-base-patch16-zero-shot":
snake_case_ :Union[str, Any] = torch.tensor([[3.50_82e-04, 9.97_85e-01, 1.79_66e-03]] )
else:
raise ValueError(f"""Model name {model_name} not supported""" )
assert torch.allclose(_lowercase, _lowercase, atol=1e-3 )
print("""Looks ok!""" )
if pytorch_dump_folder_path is not None:
print(f"""Saving model {model_name} to {pytorch_dump_folder_path}""" )
model.save_pretrained(_lowercase )
if push_to_hub:
print("""Pushing model, processor and slow tokenizer files to the hub...""" )
model.push_to_hub(_lowercase, organization="""nielsr""" )
processor.push_to_hub(_lowercase, organization="""nielsr""" )
slow_tokenizer.push_to_hub(_lowercase, organization="""nielsr""" )
if __name__ == "__main__":
__a = argparse.ArgumentParser()
# Required parameters
parser.add_argument(
"--model_name",
default="xclip-base-patch32",
type=str,
help="Name of the model.",
)
parser.add_argument(
"--pytorch_dump_folder_path", default=None, type=str, help="Path to the output PyTorch model directory."
)
parser.add_argument(
"--push_to_hub", action="store_true", help="Whether or not to push the converted model to the 🤗 hub."
)
__a = parser.parse_args()
convert_xclip_checkpoint(args.model_name, args.pytorch_dump_folder_path, args.push_to_hub)
| 66 | 0 |
"""simple docstring"""
from typing import List
from ...configuration_utils import PretrainedConfig
from ...utils import logging
UpperCAmelCase__ : Any = logging.get_logger(__name__)
UpperCAmelCase__ : str = {
'snap-research/efficientformer-l1-300': (
'https://huggingface.co/snap-research/efficientformer-l1-300/resolve/main/config.json'
),
}
class lowerCAmelCase_ (a__ ):
"""simple docstring"""
__UpperCamelCase : List[str] = '''efficientformer'''
def __init__(self , SCREAMING_SNAKE_CASE__ = [3, 2, 6, 4] , SCREAMING_SNAKE_CASE__ = [48, 96, 2_24, 4_48] , SCREAMING_SNAKE_CASE__ = [True, True, True, True] , SCREAMING_SNAKE_CASE__ = 4_48 , SCREAMING_SNAKE_CASE__ = 32 , SCREAMING_SNAKE_CASE__ = 4 , SCREAMING_SNAKE_CASE__ = 7 , SCREAMING_SNAKE_CASE__ = 5 , SCREAMING_SNAKE_CASE__ = 8 , SCREAMING_SNAKE_CASE__ = 4 , SCREAMING_SNAKE_CASE__ = 0.0 , SCREAMING_SNAKE_CASE__ = 16 , SCREAMING_SNAKE_CASE__ = 3 , SCREAMING_SNAKE_CASE__ = 3 , SCREAMING_SNAKE_CASE__ = 3 , SCREAMING_SNAKE_CASE__ = 2 , SCREAMING_SNAKE_CASE__ = 1 , SCREAMING_SNAKE_CASE__ = 0.0 , SCREAMING_SNAKE_CASE__ = 1 , SCREAMING_SNAKE_CASE__ = True , SCREAMING_SNAKE_CASE__ = True , SCREAMING_SNAKE_CASE__ = 1E-5 , SCREAMING_SNAKE_CASE__ = "gelu" , SCREAMING_SNAKE_CASE__ = 0.02 , SCREAMING_SNAKE_CASE__ = 1E-12 , SCREAMING_SNAKE_CASE__ = 2_24 , SCREAMING_SNAKE_CASE__ = 1E-05 , **SCREAMING_SNAKE_CASE__ , ) -> None:
"""simple docstring"""
super().__init__(**SCREAMING_SNAKE_CASE__ )
SCREAMING_SNAKE_CASE__ : Union[str, Any] = hidden_act
SCREAMING_SNAKE_CASE__ : Tuple = hidden_dropout_prob
SCREAMING_SNAKE_CASE__ : Tuple = hidden_sizes
SCREAMING_SNAKE_CASE__ : Tuple = num_hidden_layers
SCREAMING_SNAKE_CASE__ : Any = num_attention_heads
SCREAMING_SNAKE_CASE__ : Dict = initializer_range
SCREAMING_SNAKE_CASE__ : int = layer_norm_eps
SCREAMING_SNAKE_CASE__ : Dict = patch_size
SCREAMING_SNAKE_CASE__ : Tuple = num_channels
SCREAMING_SNAKE_CASE__ : Any = depths
SCREAMING_SNAKE_CASE__ : Dict = mlp_expansion_ratio
SCREAMING_SNAKE_CASE__ : int = downsamples
SCREAMING_SNAKE_CASE__ : Optional[int] = dim
SCREAMING_SNAKE_CASE__ : Tuple = key_dim
SCREAMING_SNAKE_CASE__ : Tuple = attention_ratio
SCREAMING_SNAKE_CASE__ : Union[str, Any] = resolution
SCREAMING_SNAKE_CASE__ : List[str] = pool_size
SCREAMING_SNAKE_CASE__ : Union[str, Any] = downsample_patch_size
SCREAMING_SNAKE_CASE__ : Union[str, Any] = downsample_stride
SCREAMING_SNAKE_CASE__ : int = downsample_pad
SCREAMING_SNAKE_CASE__ : List[str] = drop_path_rate
SCREAMING_SNAKE_CASE__ : Any = num_metaad_blocks
SCREAMING_SNAKE_CASE__ : List[Any] = distillation
SCREAMING_SNAKE_CASE__ : Tuple = use_layer_scale
SCREAMING_SNAKE_CASE__ : List[Any] = layer_scale_init_value
SCREAMING_SNAKE_CASE__ : List[Any] = image_size
SCREAMING_SNAKE_CASE__ : str = batch_norm_eps
| 25 |
"""simple docstring"""
import unittest
import numpy as np
from transformers import BertConfig, is_flax_available
from transformers.testing_utils import require_flax, slow
from ...test_modeling_flax_common import FlaxModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask
if is_flax_available():
from transformers.models.bert.modeling_flax_bert import (
FlaxBertForMaskedLM,
FlaxBertForMultipleChoice,
FlaxBertForNextSentencePrediction,
FlaxBertForPreTraining,
FlaxBertForQuestionAnswering,
FlaxBertForSequenceClassification,
FlaxBertForTokenClassification,
FlaxBertModel,
)
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def __init__( self: List[Any] , snake_case: List[str] , snake_case: Optional[Any]=13 , snake_case: List[str]=7 , snake_case: Dict=True , snake_case: List[str]=True , snake_case: Optional[int]=True , snake_case: Any=True , snake_case: Optional[Any]=99 , snake_case: Tuple=32 , snake_case: Tuple=5 , snake_case: Dict=4 , snake_case: Optional[Any]=37 , snake_case: Union[str, Any]="gelu" , snake_case: Tuple=0.1 , snake_case: List[Any]=0.1 , snake_case: List[str]=512 , snake_case: Optional[int]=16 , snake_case: int=2 , snake_case: List[Any]=0.0_2 , snake_case: Union[str, Any]=4 , ) -> List[str]:
snake_case_ :Dict = parent
snake_case_ :Any = batch_size
snake_case_ :Any = seq_length
snake_case_ :List[str] = is_training
snake_case_ :Optional[Any] = use_attention_mask
snake_case_ :Dict = use_token_type_ids
snake_case_ :Union[str, Any] = use_labels
snake_case_ :str = vocab_size
snake_case_ :int = hidden_size
snake_case_ :List[str] = num_hidden_layers
snake_case_ :Dict = num_attention_heads
snake_case_ :Any = intermediate_size
snake_case_ :Tuple = hidden_act
snake_case_ :int = hidden_dropout_prob
snake_case_ :Optional[Any] = attention_probs_dropout_prob
snake_case_ :Any = max_position_embeddings
snake_case_ :Union[str, Any] = type_vocab_size
snake_case_ :Optional[int] = type_sequence_label_size
snake_case_ :Union[str, Any] = initializer_range
snake_case_ :Tuple = num_choices
def lowerCAmelCase_ ( self: Tuple ) -> str:
snake_case_ :Optional[Any] = ids_tensor([self.batch_size, self.seq_length] , self.vocab_size )
snake_case_ :Union[str, Any] = None
if self.use_attention_mask:
snake_case_ :str = random_attention_mask([self.batch_size, self.seq_length] )
snake_case_ :Any = None
if self.use_token_type_ids:
snake_case_ :List[str] = ids_tensor([self.batch_size, self.seq_length] , self.type_vocab_size )
snake_case_ :int = BertConfig(
vocab_size=self.vocab_size , hidden_size=self.hidden_size , num_hidden_layers=self.num_hidden_layers , num_attention_heads=self.num_attention_heads , intermediate_size=self.intermediate_size , hidden_act=self.hidden_act , hidden_dropout_prob=self.hidden_dropout_prob , attention_probs_dropout_prob=self.attention_probs_dropout_prob , max_position_embeddings=self.max_position_embeddings , type_vocab_size=self.type_vocab_size , is_decoder=snake_case , initializer_range=self.initializer_range , )
return config, input_ids, token_type_ids, attention_mask
def lowerCAmelCase_ ( self: Optional[int] ) -> int:
snake_case_ :str = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_, snake_case_ :Optional[int] = config_and_inputs
snake_case_ :Union[str, Any] = {"""input_ids""": input_ids, """token_type_ids""": token_type_ids, """attention_mask""": attention_mask}
return config, inputs_dict
def lowerCAmelCase_ ( self: Optional[Any] ) -> Any:
snake_case_ :int = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_, snake_case_ :Dict = config_and_inputs
snake_case_ :Union[str, Any] = True
snake_case_ :Optional[int] = floats_tensor([self.batch_size, self.seq_length, self.hidden_size] )
snake_case_ :Tuple = ids_tensor([self.batch_size, self.seq_length] , vocab_size=2 )
return (
config,
input_ids,
attention_mask,
encoder_hidden_states,
encoder_attention_mask,
)
@require_flax
class lowerCamelCase ( _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[str] = True
_A : Dict = (
(
FlaxBertModel,
FlaxBertForPreTraining,
FlaxBertForMaskedLM,
FlaxBertForMultipleChoice,
FlaxBertForQuestionAnswering,
FlaxBertForNextSentencePrediction,
FlaxBertForSequenceClassification,
FlaxBertForTokenClassification,
FlaxBertForQuestionAnswering,
)
if is_flax_available()
else ()
)
def lowerCAmelCase_ ( self: int ) -> List[str]:
snake_case_ :Any = FlaxBertModelTester(self )
@slow
def lowerCAmelCase_ ( self: List[str] ) -> Dict:
# Only check this for base model, not necessary for all model classes.
# This will also help speed-up tests.
snake_case_ :Dict = FlaxBertModel.from_pretrained("""bert-base-cased""" )
snake_case_ :Dict = model(np.ones((1, 1) ) )
self.assertIsNotNone(snake_case )
| 66 | 0 |
_snake_case = "0.21.0"
from .accelerator import Accelerator
from .big_modeling import (
cpu_offload,
cpu_offload_with_hook,
disk_offload,
dispatch_model,
init_empty_weights,
init_on_device,
load_checkpoint_and_dispatch,
)
from .data_loader import skip_first_batches
from .launchers import debug_launcher, notebook_launcher
from .state import PartialState
from .utils import (
DeepSpeedPlugin,
DistributedDataParallelKwargs,
DistributedType,
FullyShardedDataParallelPlugin,
GradScalerKwargs,
InitProcessGroupKwargs,
find_executable_batch_size,
infer_auto_device_map,
is_rich_available,
load_checkpoint_in_model,
synchronize_rng_states,
)
if is_rich_available():
from .utils import rich
| 26 |
"""simple docstring"""
import math
class lowerCamelCase :
'''simple docstring'''
def lowerCAmelCase_ ( self: Tuple , snake_case: list[list[float]] , snake_case: list[int] ) -> int:
snake_case_ :Any = 0.0
snake_case_ :Tuple = 0.0
for i in range(len(snake_case ) ):
da += math.pow((sample[i] - weights[0][i]) , 2 )
da += math.pow((sample[i] - weights[1][i]) , 2 )
return 0 if da > da else 1
return 0
def lowerCAmelCase_ ( self: Optional[int] , snake_case: list[list[int | float]] , snake_case: list[int] , snake_case: int , snake_case: float ) -> list[list[int | float]]:
for i in range(len(snake_case ) ):
weights[j][i] += alpha * (sample[i] - weights[j][i])
return weights
def A_ ( ):
'''simple docstring'''
snake_case_ :Dict = [[1, 1, 0, 0], [0, 0, 0, 1], [1, 0, 0, 0], [0, 0, 1, 1]]
# weight initialization ( n, C )
snake_case_ :List[Any] = [[0.2, 0.6, 0.5, 0.9], [0.8, 0.4, 0.7, 0.3]]
# training
snake_case_ :Optional[Any] = SelfOrganizingMap()
snake_case_ :Dict = 3
snake_case_ :Dict = 0.5
for _ in range(_lowercase ):
for j in range(len(_lowercase ) ):
# training sample
snake_case_ :List[Any] = training_samples[j]
# Compute the winning vector
snake_case_ :Optional[int] = self_organizing_map.get_winner(_lowercase, _lowercase )
# Update the winning vector
snake_case_ :List[str] = self_organizing_map.update(_lowercase, _lowercase, _lowercase, _lowercase )
# classify test sample
snake_case_ :str = [0, 0, 0, 1]
snake_case_ :List[Any] = self_organizing_map.get_winner(_lowercase, _lowercase )
# results
print(f"""Clusters that the test sample belongs to : {winner}""" )
print(f"""Weights that have been trained : {weights}""" )
# running the main() function
if __name__ == "__main__":
main()
| 66 | 0 |
'''simple docstring'''
import torch
from transformers import AutoModel
class __UpperCamelCase ( torch.nn.Module ):
def __init__( self , __a="sayef/fsner-bert-base-uncased" ):
'''simple docstring'''
super(__a , self ).__init__()
__a : Tuple = AutoModel.from_pretrained(__a , return_dict=__a )
__a : int = torch.nn.CosineSimilarity(3 , 1E-0_8 )
__a : Union[str, Any] = torch.nn.Softmax(dim=1 )
def __UpperCAmelCase ( self , **__a ):
'''simple docstring'''
return self.bert(**__a ).last_hidden_state
def __UpperCAmelCase ( self , __a ):
'''simple docstring'''
return token_embeddings.sum(2 , keepdim=__a )
def __UpperCAmelCase ( self , __a , __a , __a=1 ):
'''simple docstring'''
return self.softmax(T * self.cos(__a , __a ) )
def __UpperCAmelCase ( self , __a , __a ):
'''simple docstring'''
__a : str = W_supports['sizes'].tolist()
__a : Union[str, Any] = W_supports['start_token_id'].item()
__a : Any = W_supports['end_token_id'].item()
del W_supports["sizes"]
del W_supports["start_token_id"]
del W_supports["end_token_id"]
__a : Tuple = self.BERT(**__a )
__a : str = self.BERT(**__a )
__a : Any = None
__a : Dict = None
__a : Dict = W_supports['input_ids'] == start_token_id
__a : Union[str, Any] = W_supports['input_ids'] == end_token_id
for i, size in enumerate(__a ):
if i == 0:
__a : Optional[int] = 0
else:
__a : Union[str, Any] = support_sizes[i - 1]
__a : int = S[s : s + size][start_token_masks[s : s + size]]
__a : Union[str, Any] = S[s : s + size][end_token_masks[s : s + size]]
__a : Tuple = torch.matmul(q[i] , s_start.T ).sum(1 ).softmax(0 )
__a : Dict = torch.matmul(q[i] , s_end.T ).sum(1 ).softmax(0 )
if p_starts is not None:
__a : str = torch.vstack((p_starts, p_start) )
__a : str = torch.vstack((p_ends, p_end) )
else:
__a : List[str] = p_start
__a : int = p_end
return p_starts, p_ends
| 27 |
"""simple docstring"""
import collections
import inspect
import unittest
from transformers import SwinvaConfig
from transformers.testing_utils import require_torch, require_vision, slow, torch_device
from transformers.utils import cached_property, is_torch_available, is_vision_available
from ...test_configuration_common import ConfigTester
from ...test_modeling_common import ModelTesterMixin, _config_zero_init, floats_tensor, ids_tensor
from ...test_pipeline_mixin import PipelineTesterMixin
if is_torch_available():
import torch
from torch import nn
from transformers import SwinvaForImageClassification, SwinvaForMaskedImageModeling, SwinvaModel
from transformers.models.swinva.modeling_swinva import SWINV2_PRETRAINED_MODEL_ARCHIVE_LIST
if is_vision_available():
from PIL import Image
from transformers import AutoImageProcessor
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Optional[int] , snake_case: Any , snake_case: Optional[Any]=13 , snake_case: Tuple=32 , snake_case: Optional[int]=2 , snake_case: Tuple=3 , snake_case: Tuple=16 , snake_case: Optional[Any]=[1, 2, 1] , snake_case: Optional[int]=[2, 2, 4] , snake_case: Optional[int]=2 , snake_case: int=2.0 , snake_case: Union[str, Any]=True , snake_case: List[str]=0.0 , snake_case: List[Any]=0.0 , snake_case: Optional[Any]=0.1 , snake_case: List[Any]="gelu" , snake_case: Optional[int]=False , snake_case: Union[str, Any]=True , snake_case: Union[str, Any]=0.0_2 , snake_case: Optional[int]=1E-5 , snake_case: Optional[Any]=True , snake_case: List[Any]=None , snake_case: List[Any]=True , snake_case: Optional[Any]=10 , snake_case: str=8 , ) -> Tuple:
snake_case_ :Dict = parent
snake_case_ :Any = batch_size
snake_case_ :List[Any] = image_size
snake_case_ :List[Any] = patch_size
snake_case_ :int = num_channels
snake_case_ :Tuple = embed_dim
snake_case_ :str = depths
snake_case_ :str = num_heads
snake_case_ :Optional[int] = window_size
snake_case_ :Tuple = mlp_ratio
snake_case_ :Any = qkv_bias
snake_case_ :List[Any] = hidden_dropout_prob
snake_case_ :Optional[Any] = attention_probs_dropout_prob
snake_case_ :Union[str, Any] = drop_path_rate
snake_case_ :Any = hidden_act
snake_case_ :Optional[Any] = use_absolute_embeddings
snake_case_ :Union[str, Any] = patch_norm
snake_case_ :Dict = layer_norm_eps
snake_case_ :str = initializer_range
snake_case_ :Tuple = is_training
snake_case_ :Tuple = scope
snake_case_ :Union[str, Any] = use_labels
snake_case_ :Optional[Any] = type_sequence_label_size
snake_case_ :Dict = encoder_stride
def lowerCAmelCase_ ( self: int ) -> int:
snake_case_ :List[str] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case_ :Any = None
if self.use_labels:
snake_case_ :str = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ :int = self.get_config()
return config, pixel_values, labels
def lowerCAmelCase_ ( self: str ) -> Union[str, Any]:
return SwinvaConfig(
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 , )
def lowerCAmelCase_ ( self: str , snake_case: Optional[int] , snake_case: Dict , snake_case: str ) -> List[Any]:
snake_case_ :Union[str, Any] = SwinvaModel(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Optional[int] = model(snake_case )
snake_case_ :Optional[Any] = ((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 lowerCAmelCase_ ( self: int , snake_case: List[str] , snake_case: Tuple , snake_case: int ) -> Any:
snake_case_ :Dict = SwinvaForMaskedImageModeling(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Tuple = model(snake_case )
self.parent.assertEqual(
result.logits.shape , (self.batch_size, self.num_channels, self.image_size, self.image_size) )
# test greyscale images
snake_case_ :List[Any] = 1
snake_case_ :int = SwinvaForMaskedImageModeling(snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Union[str, Any] = floats_tensor([self.batch_size, 1, self.image_size, self.image_size] )
snake_case_ :int = model(snake_case )
self.parent.assertEqual(result.logits.shape , (self.batch_size, 1, self.image_size, self.image_size) )
def lowerCAmelCase_ ( self: List[Any] , snake_case: Any , snake_case: List[str] , snake_case: Union[str, Any] ) -> Tuple:
snake_case_ :int = self.type_sequence_label_size
snake_case_ :List[Any] = SwinvaForImageClassification(snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Dict = model(snake_case , labels=snake_case )
self.parent.assertEqual(result.logits.shape , (self.batch_size, self.type_sequence_label_size) )
def lowerCAmelCase_ ( self: int ) -> str:
snake_case_ :Any = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_ :List[str] = config_and_inputs
snake_case_ :List[Any] = {"""pixel_values""": pixel_values}
return config, inputs_dict
@require_torch
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : Optional[Any] = (
(SwinvaModel, SwinvaForImageClassification, SwinvaForMaskedImageModeling) if is_torch_available() else ()
)
_A : Any = (
{"""feature-extraction""": SwinvaModel, """image-classification""": SwinvaForImageClassification}
if is_torch_available()
else {}
)
_A : List[Any] = False
_A : List[str] = False
_A : Tuple = False
_A : List[str] = False
def lowerCAmelCase_ ( self: Dict ) -> List[Any]:
snake_case_ :Optional[int] = SwinvaModelTester(self )
snake_case_ :List[str] = ConfigTester(self , config_class=snake_case , embed_dim=37 )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> List[Any]:
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 lowerCAmelCase_ ( self: Union[str, Any] ) -> Tuple:
snake_case_ :List[str] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case )
@unittest.skip(reason="""Got `CUDA error: misaligned address` with PyTorch 2.0.0.""" )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> str:
pass
@unittest.skip(reason="""Swinv2 does not use inputs_embeds""" )
def lowerCAmelCase_ ( self: int ) -> Dict:
pass
def lowerCAmelCase_ ( self: List[str] ) -> Union[str, Any]:
snake_case_, snake_case_ :List[str] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :Optional[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 lowerCAmelCase_ ( self: Dict ) -> Optional[int]:
snake_case_, snake_case_ :Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :Optional[int] = model_class(snake_case )
snake_case_ :List[Any] = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case_ :int = [*signature.parameters.keys()]
snake_case_ :List[Any] = ["""pixel_values"""]
self.assertListEqual(arg_names[:1] , snake_case )
def lowerCAmelCase_ ( self: List[str] ) -> Optional[Any]:
snake_case_, snake_case_ :List[str] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :List[str] = True
for model_class in self.all_model_classes:
snake_case_ :List[Any] = True
snake_case_ :Any = False
snake_case_ :Optional[int] = True
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_ :str = outputs.attentions
snake_case_ :Dict = len(self.model_tester.depths )
self.assertEqual(len(snake_case ) , snake_case )
# check that output_attentions also work using config
del inputs_dict["output_attentions"]
snake_case_ :Union[str, Any] = True
snake_case_ :Tuple = config.window_size**2
snake_case_ :Any = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :Union[str, Any] = model(**self._prepare_for_class(snake_case , snake_case ) )
snake_case_ :int = outputs.attentions
self.assertEqual(len(snake_case ) , snake_case )
self.assertListEqual(
list(attentions[0].shape[-3:] ) , [self.model_tester.num_heads[0], window_size_squared, window_size_squared] , )
snake_case_ :Any = len(snake_case )
# Check attention is always last and order is fine
snake_case_ :int = True
snake_case_ :Dict = True
snake_case_ :Optional[int] = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :Dict = model(**self._prepare_for_class(snake_case , snake_case ) )
if hasattr(self.model_tester , """num_hidden_states_types""" ):
snake_case_ :Any = self.model_tester.num_hidden_states_types
else:
# also another +1 for reshaped_hidden_states
snake_case_ :int = 2
self.assertEqual(out_len + added_hidden_states , len(snake_case ) )
snake_case_ :str = outputs.attentions
self.assertEqual(len(snake_case ) , snake_case )
self.assertListEqual(
list(self_attentions[0].shape[-3:] ) , [self.model_tester.num_heads[0], window_size_squared, window_size_squared] , )
def lowerCAmelCase_ ( self: int , snake_case: Dict , snake_case: Dict , snake_case: Optional[Any] , snake_case: Dict ) -> List[str]:
snake_case_ :Dict = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :Optional[int] = model(**self._prepare_for_class(snake_case , snake_case ) )
snake_case_ :str = outputs.hidden_states
snake_case_ :List[Any] = getattr(
self.model_tester , """expected_num_hidden_layers""" , len(self.model_tester.depths ) + 1 )
self.assertEqual(len(snake_case ) , snake_case )
# Swinv2 has a different seq_length
snake_case_ :List[Any] = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :Optional[int] = (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] , )
snake_case_ :str = outputs.reshaped_hidden_states
self.assertEqual(len(snake_case ) , snake_case )
snake_case_, snake_case_, snake_case_, snake_case_ :Any = reshaped_hidden_states[0].shape
snake_case_ :int = (
reshaped_hidden_states[0].view(snake_case , snake_case , height * width ).permute(0 , 2 , 1 )
)
self.assertListEqual(
list(reshaped_hidden_states.shape[-2:] ) , [num_patches, self.model_tester.embed_dim] , )
def lowerCAmelCase_ ( self: Any ) -> Any:
snake_case_, snake_case_ :List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Union[str, Any] = (
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_ :Union[str, Any] = 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_ :List[str] = True
self.check_hidden_states_output(snake_case , snake_case , snake_case , snake_case )
def lowerCAmelCase_ ( self: Tuple ) -> Any:
snake_case_, snake_case_ :Union[str, Any] = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Optional[int] = 3
snake_case_ :Union[str, Any] = (
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_ :str = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :Any = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
snake_case_ :int = 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_ :Tuple = True
self.check_hidden_states_output(snake_case , snake_case , snake_case , (padded_height, padded_width) )
def lowerCAmelCase_ ( self: Any ) -> Tuple:
snake_case_ :int = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_masked_image_modeling(*snake_case )
def lowerCAmelCase_ ( self: Optional[int] ) -> Dict:
snake_case_ :Any = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_image_classification(*snake_case )
@slow
def lowerCAmelCase_ ( self: List[Any] ) -> Dict:
for model_name in SWINV2_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:
snake_case_ :List[str] = SwinvaModel.from_pretrained(snake_case )
self.assertIsNotNone(snake_case )
def lowerCAmelCase_ ( self: Optional[int] ) -> List[Any]:
snake_case_, snake_case_ :str = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Optional[int] = _config_zero_init(snake_case )
for model_class in self.all_model_classes:
snake_case_ :Tuple = model_class(config=snake_case )
for name, param in model.named_parameters():
if "embeddings" not in name and "logit_scale" not in name and param.requires_grad:
self.assertIn(
((param.data.mean() * 1E9).round() / 1E9).item() , [0.0, 1.0] , msg=f"""Parameter {name} of model {model_class} seems not properly initialized""" , )
@require_vision
@require_torch
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
@cached_property
def lowerCAmelCase_ ( self: Optional[int] ) -> List[Any]:
return (
AutoImageProcessor.from_pretrained("""microsoft/swinv2-tiny-patch4-window8-256""" )
if is_vision_available()
else None
)
@slow
def lowerCAmelCase_ ( self: List[str] ) -> List[str]:
snake_case_ :Tuple = SwinvaForImageClassification.from_pretrained("""microsoft/swinv2-tiny-patch4-window8-256""" ).to(
snake_case )
snake_case_ :str = self.default_image_processor
snake_case_ :List[Any] = Image.open("""./tests/fixtures/tests_samples/COCO/000000039769.png""" )
snake_case_ :str = image_processor(images=snake_case , return_tensors="""pt""" ).to(snake_case )
# forward pass
with torch.no_grad():
snake_case_ :Tuple = model(**snake_case )
# verify the logits
snake_case_ :Dict = torch.Size((1, 1_000) )
self.assertEqual(outputs.logits.shape , snake_case )
snake_case_ :int = torch.tensor([-0.3_9_4_7, -0.4_3_0_6, 0.0_0_2_6] ).to(snake_case )
self.assertTrue(torch.allclose(outputs.logits[0, :3] , snake_case , atol=1E-4 ) )
| 66 | 0 |
'''simple docstring'''
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 : str = logging.get_logger(__name__)
_lowerCamelCase : Union[str, Any] = {
"Helsinki-NLP/opus-mt-en-de": "https://huggingface.co/Helsinki-NLP/opus-mt-en-de/resolve/main/config.json",
# See all Marian models at https://huggingface.co/models?filter=marian
}
class SCREAMING_SNAKE_CASE ( _a ):
"""simple docstring"""
_SCREAMING_SNAKE_CASE = """marian"""
_SCREAMING_SNAKE_CASE = ["""past_key_values"""]
_SCREAMING_SNAKE_CASE = {"""num_attention_heads""": """encoder_attention_heads""", """hidden_size""": """d_model"""}
def __init__( self : List[str] , UpperCamelCase__ : Optional[Any]=5_8_1_0_1 , UpperCamelCase__ : Dict=None , UpperCamelCase__ : Any=1_0_2_4 , UpperCamelCase__ : List[str]=1_2 , UpperCamelCase__ : Dict=4_0_9_6 , UpperCamelCase__ : Tuple=1_6 , UpperCamelCase__ : Any=1_2 , UpperCamelCase__ : List[str]=4_0_9_6 , UpperCamelCase__ : int=1_6 , UpperCamelCase__ : str=0.0 , UpperCamelCase__ : Optional[Any]=0.0 , UpperCamelCase__ : Optional[int]=True , UpperCamelCase__ : List[Any]=True , UpperCamelCase__ : Optional[int]="gelu" , UpperCamelCase__ : List[Any]=1_0_2_4 , UpperCamelCase__ : str=0.1 , UpperCamelCase__ : Dict=0.0 , UpperCamelCase__ : Tuple=0.0 , UpperCamelCase__ : List[str]=0.0_2 , UpperCamelCase__ : List[str]=5_8_1_0_0 , UpperCamelCase__ : Dict=False , UpperCamelCase__ : Optional[int]=5_8_1_0_0 , UpperCamelCase__ : Union[str, Any]=0 , UpperCamelCase__ : List[str]=0 , UpperCamelCase__ : int=True , **UpperCamelCase__ : Optional[Any] , ):
"""simple docstring"""
UpperCamelCase = vocab_size
UpperCamelCase = decoder_vocab_size or vocab_size
UpperCamelCase = max_position_embeddings
UpperCamelCase = d_model
UpperCamelCase = encoder_ffn_dim
UpperCamelCase = encoder_layers
UpperCamelCase = encoder_attention_heads
UpperCamelCase = decoder_ffn_dim
UpperCamelCase = decoder_layers
UpperCamelCase = decoder_attention_heads
UpperCamelCase = dropout
UpperCamelCase = attention_dropout
UpperCamelCase = activation_dropout
UpperCamelCase = activation_function
UpperCamelCase = init_std
UpperCamelCase = encoder_layerdrop
UpperCamelCase = decoder_layerdrop
UpperCamelCase = use_cache
UpperCamelCase = encoder_layers
UpperCamelCase = scale_embedding # scale factor will be sqrt(d_model) if True
UpperCamelCase = share_encoder_decoder_embeddings
super().__init__(
pad_token_id=UpperCamelCase__ , eos_token_id=UpperCamelCase__ , is_encoder_decoder=UpperCamelCase__ , decoder_start_token_id=UpperCamelCase__ , forced_eos_token_id=UpperCamelCase__ , **UpperCamelCase__ , )
class SCREAMING_SNAKE_CASE ( _a ):
"""simple docstring"""
@property
# Copied from transformers.models.bart.configuration_bart.BartOnnxConfig.inputs
def A ( self : List[str] ):
"""simple docstring"""
if self.task in ["default", "seq2seq-lm"]:
UpperCamelCase = OrderedDict(
[
('input_ids', {0: 'batch', 1: 'encoder_sequence'}),
('attention_mask', {0: 'batch', 1: 'encoder_sequence'}),
] )
if self.use_past:
UpperCamelCase = {0: 'batch'}
UpperCamelCase = {0: 'batch', 1: 'past_decoder_sequence + sequence'}
else:
UpperCamelCase = {0: 'batch', 1: 'decoder_sequence'}
UpperCamelCase = {0: 'batch', 1: 'decoder_sequence'}
if self.use_past:
self.fill_with_past_key_values_(UpperCamelCase__ , direction='inputs' )
elif self.task == "causal-lm":
# TODO: figure this case out.
UpperCamelCase = OrderedDict(
[
('input_ids', {0: 'batch', 1: 'encoder_sequence'}),
('attention_mask', {0: 'batch', 1: 'encoder_sequence'}),
] )
if self.use_past:
UpperCamelCase , UpperCamelCase = self.num_layers
for i in range(UpperCamelCase__ ):
UpperCamelCase = {0: 'batch', 2: 'past_sequence + sequence'}
UpperCamelCase = {0: 'batch', 2: 'past_sequence + sequence'}
else:
UpperCamelCase = 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
# Copied from transformers.models.bart.configuration_bart.BartOnnxConfig.outputs
def A ( self : Dict ):
"""simple docstring"""
if self.task in ["default", "seq2seq-lm"]:
UpperCamelCase = super().outputs
else:
UpperCamelCase = super(UpperCamelCase__ , self ).outputs
if self.use_past:
UpperCamelCase , UpperCamelCase = self.num_layers
for i in range(UpperCamelCase__ ):
UpperCamelCase = {0: 'batch', 2: 'past_sequence + sequence'}
UpperCamelCase = {0: 'batch', 2: 'past_sequence + sequence'}
return common_outputs
def A ( self : List[str] , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ):
"""simple docstring"""
UpperCamelCase = self._generate_dummy_inputs_for_encoder_and_decoder(
UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ )
# Generate decoder inputs
UpperCamelCase = seq_length if not self.use_past else 1
UpperCamelCase = self._generate_dummy_inputs_for_encoder_and_decoder(
UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ )
UpperCamelCase = {f"""decoder_{name}""": tensor for name, tensor in decoder_inputs.items()}
UpperCamelCase = dict(**UpperCamelCase__ , **UpperCamelCase__ )
if self.use_past:
if not is_torch_available():
raise ValueError('Cannot generate dummy past_keys inputs without PyTorch installed.' )
else:
import torch
UpperCamelCase , UpperCamelCase = common_inputs['input_ids'].shape
UpperCamelCase = common_inputs['decoder_input_ids'].shape[1]
UpperCamelCase , UpperCamelCase = self.num_attention_heads
UpperCamelCase = (
batch,
num_encoder_attention_heads,
encoder_seq_length,
self._config.hidden_size // num_encoder_attention_heads,
)
UpperCamelCase = decoder_seq_length + 3
UpperCamelCase = (
batch,
num_decoder_attention_heads,
decoder_past_length,
self._config.hidden_size // num_decoder_attention_heads,
)
UpperCamelCase = torch.cat(
[common_inputs['decoder_attention_mask'], torch.ones(UpperCamelCase__ , UpperCamelCase__ )] , dim=1 )
UpperCamelCase = []
# If the number of encoder and decoder layers are present in the model configuration, both are considered
UpperCamelCase , UpperCamelCase = self.num_layers
UpperCamelCase = min(UpperCamelCase__ , UpperCamelCase__ )
UpperCamelCase = max(UpperCamelCase__ , UpperCamelCase__ ) - min_num_layers
UpperCamelCase = 'encoder' if num_encoder_layers > num_decoder_layers else 'decoder'
for _ in range(UpperCamelCase__ ):
common_inputs["past_key_values"].append(
(
torch.zeros(UpperCamelCase__ ),
torch.zeros(UpperCamelCase__ ),
torch.zeros(UpperCamelCase__ ),
torch.zeros(UpperCamelCase__ ),
) )
# TODO: test this.
UpperCamelCase = encoder_shape if remaining_side_name == 'encoder' else decoder_shape
for _ in range(UpperCamelCase__ , UpperCamelCase__ ):
common_inputs["past_key_values"].append((torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ )) )
return common_inputs
def A ( self : Tuple , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ):
"""simple docstring"""
UpperCamelCase = self._generate_dummy_inputs_for_encoder_and_decoder(
UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ )
if self.use_past:
if not is_torch_available():
raise ValueError('Cannot generate dummy past_keys inputs without PyTorch installed.' )
else:
import torch
UpperCamelCase , UpperCamelCase = common_inputs['input_ids'].shape
# Not using the same length for past_key_values
UpperCamelCase = seqlen + 2
UpperCamelCase , UpperCamelCase = self.num_layers
UpperCamelCase , UpperCamelCase = self.num_attention_heads
UpperCamelCase = (
batch,
num_encoder_attention_heads,
past_key_values_length,
self._config.hidden_size // num_encoder_attention_heads,
)
UpperCamelCase = common_inputs['attention_mask'].dtype
UpperCamelCase = torch.cat(
[common_inputs['attention_mask'], torch.ones(UpperCamelCase__ , UpperCamelCase__ , dtype=UpperCamelCase__ )] , dim=1 )
UpperCamelCase = [
(torch.zeros(UpperCamelCase__ ), torch.zeros(UpperCamelCase__ )) for _ in range(UpperCamelCase__ )
]
return common_inputs
def A ( self : int , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ):
"""simple docstring"""
UpperCamelCase = compute_effective_axis_dimension(
UpperCamelCase__ , fixed_dimension=OnnxConfig.default_fixed_batch , num_token_to_add=0 )
# If dynamic axis (-1) we forward with a fixed dimension of 8 tokens to avoid optimizations made by ONNX
UpperCamelCase = tokenizer.num_special_tokens_to_add(UpperCamelCase__ )
UpperCamelCase = compute_effective_axis_dimension(
UpperCamelCase__ , fixed_dimension=OnnxConfig.default_fixed_sequence , num_token_to_add=UpperCamelCase__ )
# Generate dummy inputs according to compute batch and sequence
UpperCamelCase = [' '.join([tokenizer.unk_token] ) * seq_length] * batch_size
UpperCamelCase = dict(tokenizer(UpperCamelCase__ , return_tensors=UpperCamelCase__ ) )
return common_inputs
def A ( self : Dict , UpperCamelCase__ : PreTrainedTokenizer , UpperCamelCase__ : int = -1 , UpperCamelCase__ : int = -1 , UpperCamelCase__ : bool = False , UpperCamelCase__ : Optional[TensorType] = None , ):
"""simple docstring"""
if self.task in ["default", "seq2seq-lm"]:
UpperCamelCase = self._generate_dummy_inputs_for_default_and_seqaseq_lm(
UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ )
else:
UpperCamelCase = self._generate_dummy_inputs_for_causal_lm(
UpperCamelCase__ , batch_size=UpperCamelCase__ , seq_length=UpperCamelCase__ , is_pair=UpperCamelCase__ , framework=UpperCamelCase__ )
return common_inputs
def A ( self : List[str] , UpperCamelCase__ : Any , UpperCamelCase__ : List[str] , UpperCamelCase__ : Tuple , UpperCamelCase__ : Optional[int] ):
"""simple docstring"""
if self.task in ["default", "seq2seq-lm"]:
UpperCamelCase = super()._flatten_past_key_values_(UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ )
else:
UpperCamelCase = super(UpperCamelCase__ , self )._flatten_past_key_values_(
UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ , UpperCamelCase__ )
@property
def A ( self : Dict ):
"""simple docstring"""
return 1E-4
| 28 |
"""simple docstring"""
import re
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Optional[int] = re.compile(
r"""^(?:0|94|\+94|0{2}94)""" r"""7(0|1|2|4|5|6|7|8)""" r"""(-| |)""" r"""\d{7}$""" )
return bool(re.search(_lowercase, _lowercase ) )
if __name__ == "__main__":
__a = "0094702343221"
print(is_sri_lankan_phone_number(phone))
| 66 | 0 |
import heapq
def lowercase__ ( __snake_case : dict ):
'''simple docstring'''
UpperCAmelCase_ : list[list] = []
# for each node and his adjacency list add them and the rank of the node to queue
# using heapq module the queue will be filled like a Priority Queue
# heapq works with a min priority queue, so I used -1*len(v) to build it
for key, value in graph.items():
# O(log(n))
heapq.heappush(__snake_case , [-1 * len(__snake_case ), (key, value)] )
# chosen_vertices = set of chosen vertices
UpperCAmelCase_ : List[str] = set()
# while queue isn't empty and there are still edges
# (queue[0][0] is the rank of the node with max rank)
while queue and queue[0][0] != 0:
# extract vertex with max rank from queue and add it to chosen_vertices
UpperCAmelCase_ : Any = heapq.heappop(__snake_case )[1][0]
chosen_vertices.add(__snake_case )
# Remove all arcs adjacent to argmax
for elem in queue:
# if v haven't adjacent node, skip
if elem[0] == 0:
continue
# if argmax is reachable from elem
# remove argmax from elem's adjacent list and update his rank
if argmax in elem[1][1]:
UpperCAmelCase_ : Tuple = elem[1][1].index(__snake_case )
del elem[1][1][index]
elem[0] += 1
# re-order the queue
heapq.heapify(__snake_case )
return chosen_vertices
if __name__ == "__main__":
import doctest
doctest.testmod()
__UpperCAmelCase = {0: [1, 3], 1: [0, 3], 2: [0, 3, 4], 3: [0, 1, 2], 4: [2, 3]}
print(F'Minimum vertex cover:\n{greedy_min_vertex_cover(graph)}')
| 29 |
"""simple docstring"""
import argparse
import json
import os
import pickle
import shutil
import numpy as np
import torch
from distiller import Distiller
from lm_seqs_dataset import LmSeqsDataset
from transformers import (
BertConfig,
BertForMaskedLM,
BertTokenizer,
DistilBertConfig,
DistilBertForMaskedLM,
DistilBertTokenizer,
GPTaConfig,
GPTaLMHeadModel,
GPTaTokenizer,
RobertaConfig,
RobertaForMaskedLM,
RobertaTokenizer,
)
from utils import git_log, init_gpu_params, logger, set_seed
__a = {
"distilbert": (DistilBertConfig, DistilBertForMaskedLM, DistilBertTokenizer),
"roberta": (RobertaConfig, RobertaForMaskedLM, RobertaTokenizer),
"bert": (BertConfig, BertForMaskedLM, BertTokenizer),
"gpt2": (GPTaConfig, GPTaLMHeadModel, GPTaTokenizer),
}
def A_ ( _lowercase ):
'''simple docstring'''
assert (args.mlm and args.alpha_mlm > 0.0) or (not args.mlm and args.alpha_mlm == 0.0)
assert (args.alpha_mlm > 0.0 and args.alpha_clm == 0.0) or (args.alpha_mlm == 0.0 and args.alpha_clm > 0.0)
if args.mlm:
assert os.path.isfile(args.token_counts )
assert (args.student_type in ["roberta", "distilbert"]) and (args.teacher_type in ["roberta", "bert"])
else:
assert (args.student_type in ["gpt2"]) and (args.teacher_type in ["gpt2"])
assert args.teacher_type == args.student_type or (
args.student_type == "distilbert" and args.teacher_type == "bert"
)
assert os.path.isfile(args.student_config )
if args.student_pretrained_weights is not None:
assert os.path.isfile(args.student_pretrained_weights )
if args.freeze_token_type_embds:
assert args.student_type in ["roberta"]
assert args.alpha_ce >= 0.0
assert args.alpha_mlm >= 0.0
assert args.alpha_clm >= 0.0
assert args.alpha_mse >= 0.0
assert args.alpha_cos >= 0.0
assert args.alpha_ce + args.alpha_mlm + args.alpha_clm + args.alpha_mse + args.alpha_cos > 0.0
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
if args.student_type == "roberta":
snake_case_ :Tuple = False
elif args.student_type == "gpt2":
snake_case_ :Union[str, Any] = False
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
if args.student_type == "roberta":
snake_case_ :List[str] = False
def A_ ( ):
'''simple docstring'''
snake_case_ :Union[str, Any] = argparse.ArgumentParser(description="""Training""" )
parser.add_argument("""--force""", action="""store_true""", help="""Overwrite dump_path if it already exists.""" )
parser.add_argument(
"""--dump_path""", type=_lowercase, required=_lowercase, help="""The output directory (log, checkpoints, parameters, etc.)""" )
parser.add_argument(
"""--data_file""", type=_lowercase, required=_lowercase, help="""The binarized file (tokenized + tokens_to_ids) and grouped by sequence.""", )
parser.add_argument(
"""--student_type""", type=_lowercase, choices=["""distilbert""", """roberta""", """gpt2"""], required=_lowercase, help="""The student type (DistilBERT, RoBERTa).""", )
parser.add_argument("""--student_config""", type=_lowercase, required=_lowercase, help="""Path to the student configuration.""" )
parser.add_argument(
"""--student_pretrained_weights""", default=_lowercase, type=_lowercase, help="""Load student initialization checkpoint.""" )
parser.add_argument(
"""--teacher_type""", choices=["""bert""", """roberta""", """gpt2"""], required=_lowercase, help="""Teacher type (BERT, RoBERTa).""" )
parser.add_argument("""--teacher_name""", type=_lowercase, required=_lowercase, help="""The teacher model.""" )
parser.add_argument("""--temperature""", default=2.0, type=_lowercase, help="""Temperature for the softmax temperature.""" )
parser.add_argument(
"""--alpha_ce""", default=0.5, type=_lowercase, help="""Linear weight for the distillation loss. Must be >=0.""" )
parser.add_argument(
"""--alpha_mlm""", default=0.0, type=_lowercase, help="""Linear weight for the MLM loss. Must be >=0. Should be used in conjunction with `mlm` flag.""", )
parser.add_argument("""--alpha_clm""", default=0.5, type=_lowercase, help="""Linear weight for the CLM loss. Must be >=0.""" )
parser.add_argument("""--alpha_mse""", default=0.0, type=_lowercase, help="""Linear weight of the MSE loss. Must be >=0.""" )
parser.add_argument(
"""--alpha_cos""", default=0.0, type=_lowercase, help="""Linear weight of the cosine embedding loss. Must be >=0.""" )
parser.add_argument(
"""--mlm""", action="""store_true""", help="""The LM step: MLM or CLM. If `mlm` is True, the MLM is used over CLM.""" )
parser.add_argument(
"""--mlm_mask_prop""", default=0.15, type=_lowercase, help="""Proportion of tokens for which we need to make a prediction.""", )
parser.add_argument("""--word_mask""", default=0.8, type=_lowercase, help="""Proportion of tokens to mask out.""" )
parser.add_argument("""--word_keep""", default=0.1, type=_lowercase, help="""Proportion of tokens to keep.""" )
parser.add_argument("""--word_rand""", default=0.1, type=_lowercase, help="""Proportion of tokens to randomly replace.""" )
parser.add_argument(
"""--mlm_smoothing""", default=0.7, type=_lowercase, help="""Smoothing parameter to emphasize more rare tokens (see XLM, similar to word2vec).""", )
parser.add_argument("""--token_counts""", type=_lowercase, help="""The token counts in the data_file for MLM.""" )
parser.add_argument(
"""--restrict_ce_to_mask""", action="""store_true""", help="""If true, compute the distillation loss only the [MLM] prediction distribution.""", )
parser.add_argument(
"""--freeze_pos_embs""", action="""store_true""", help="""Freeze positional embeddings during distillation. For student_type in ['roberta', 'gpt2'] only.""", )
parser.add_argument(
"""--freeze_token_type_embds""", action="""store_true""", help="""Freeze token type embeddings during distillation if existent. For student_type in ['roberta'] only.""", )
parser.add_argument("""--n_epoch""", type=_lowercase, default=3, help="""Number of pass on the whole dataset.""" )
parser.add_argument("""--batch_size""", type=_lowercase, default=5, help="""Batch size (for each process).""" )
parser.add_argument(
"""--group_by_size""", action="""store_false""", help="""If true, group sequences that have similar length into the same batch. Default is true.""", )
parser.add_argument(
"""--gradient_accumulation_steps""", type=_lowercase, default=50, help="""Gradient accumulation for larger training batches.""", )
parser.add_argument("""--warmup_prop""", default=0.05, type=_lowercase, help="""Linear warmup proportion.""" )
parser.add_argument("""--weight_decay""", default=0.0, type=_lowercase, help="""Weight decay if we apply some.""" )
parser.add_argument("""--learning_rate""", default=5e-4, type=_lowercase, help="""The initial learning rate for Adam.""" )
parser.add_argument("""--adam_epsilon""", default=1e-6, type=_lowercase, help="""Epsilon for Adam optimizer.""" )
parser.add_argument("""--max_grad_norm""", default=5.0, type=_lowercase, help="""Max gradient norm.""" )
parser.add_argument("""--initializer_range""", default=0.02, type=_lowercase, help="""Random initialization range.""" )
parser.add_argument(
"""--fp16""", action="""store_true""", help="""Whether to use 16-bit (mixed) precision (through NVIDIA apex) instead of 32-bit""", )
parser.add_argument(
"""--fp16_opt_level""", type=_lowercase, default="""O1""", help=(
"""For fp16: Apex AMP optimization level selected in ['O0', 'O1', 'O2', and 'O3']."""
"""See details at https://nvidia.github.io/apex/amp.html"""
), )
parser.add_argument("""--n_gpu""", type=_lowercase, default=1, help="""Number of GPUs in the node.""" )
parser.add_argument("""--local_rank""", type=_lowercase, default=-1, help="""Distributed training - Local rank""" )
parser.add_argument("""--seed""", type=_lowercase, default=56, help="""Random seed""" )
parser.add_argument("""--log_interval""", type=_lowercase, default=500, help="""Tensorboard logging interval.""" )
parser.add_argument("""--checkpoint_interval""", type=_lowercase, default=4000, help="""Checkpoint interval.""" )
snake_case_ :Tuple = parser.parse_args()
sanity_checks(_lowercase )
# ARGS #
init_gpu_params(_lowercase )
set_seed(_lowercase )
if args.is_master:
if os.path.exists(args.dump_path ):
if not args.force:
raise ValueError(
f"""Serialization dir {args.dump_path} already exists, but you have not precised wheter to overwrite"""
""" itUse `--force` if you want to overwrite it""" )
else:
shutil.rmtree(args.dump_path )
if not os.path.exists(args.dump_path ):
os.makedirs(args.dump_path )
logger.info(f"""Experiment will be dumped and logged in {args.dump_path}""" )
# SAVE PARAMS #
logger.info(f"""Param: {args}""" )
with open(os.path.join(args.dump_path, """parameters.json""" ), """w""" ) as f:
json.dump(vars(_lowercase ), _lowercase, indent=4 )
git_log(args.dump_path )
snake_case_, snake_case_, snake_case_ :Any = MODEL_CLASSES[args.student_type]
snake_case_, snake_case_, snake_case_ :int = MODEL_CLASSES[args.teacher_type]
# TOKENIZER #
snake_case_ :Any = teacher_tokenizer_class.from_pretrained(args.teacher_name )
snake_case_ :Optional[Any] = {}
for tok_name, tok_symbol in tokenizer.special_tokens_map.items():
snake_case_ :Union[str, Any] = tokenizer.all_special_tokens.index(_lowercase )
snake_case_ :Union[str, Any] = tokenizer.all_special_ids[idx]
logger.info(f"""Special tokens {special_tok_ids}""" )
snake_case_ :str = special_tok_ids
snake_case_ :Any = tokenizer.max_model_input_sizes[args.teacher_name]
# DATA LOADER #
logger.info(f"""Loading data from {args.data_file}""" )
with open(args.data_file, """rb""" ) as fp:
snake_case_ :str = pickle.load(_lowercase )
if args.mlm:
logger.info(f"""Loading token counts from {args.token_counts} (already pre-computed)""" )
with open(args.token_counts, """rb""" ) as fp:
snake_case_ :Optional[Any] = pickle.load(_lowercase )
snake_case_ :Tuple = np.maximum(_lowercase, 1 ) ** -args.mlm_smoothing
for idx in special_tok_ids.values():
snake_case_ :Optional[int] = 0.0 # do not predict special tokens
snake_case_ :int = torch.from_numpy(_lowercase )
else:
snake_case_ :List[str] = None
snake_case_ :Optional[int] = LmSeqsDataset(params=_lowercase, data=_lowercase )
logger.info("""Data loader created.""" )
# STUDENT #
logger.info(f"""Loading student config from {args.student_config}""" )
snake_case_ :List[Any] = student_config_class.from_pretrained(args.student_config )
snake_case_ :Union[str, Any] = True
if args.student_pretrained_weights is not None:
logger.info(f"""Loading pretrained weights from {args.student_pretrained_weights}""" )
snake_case_ :List[str] = student_model_class.from_pretrained(args.student_pretrained_weights, config=_lowercase )
else:
snake_case_ :Optional[int] = student_model_class(_lowercase )
if args.n_gpu > 0:
student.to(f"""cuda:{args.local_rank}""" )
logger.info("""Student loaded.""" )
# TEACHER #
snake_case_ :Dict = teacher_model_class.from_pretrained(args.teacher_name, output_hidden_states=_lowercase )
if args.n_gpu > 0:
teacher.to(f"""cuda:{args.local_rank}""" )
logger.info(f"""Teacher loaded from {args.teacher_name}.""" )
# FREEZING #
if args.freeze_pos_embs:
freeze_pos_embeddings(_lowercase, _lowercase )
if args.freeze_token_type_embds:
freeze_token_type_embeddings(_lowercase, _lowercase )
# SANITY CHECKS #
assert student.config.vocab_size == teacher.config.vocab_size
assert student.config.hidden_size == teacher.config.hidden_size
assert student.config.max_position_embeddings == teacher.config.max_position_embeddings
if args.mlm:
assert token_probs.size(0 ) == stu_architecture_config.vocab_size
# DISTILLER #
torch.cuda.empty_cache()
snake_case_ :Optional[int] = Distiller(
params=_lowercase, dataset=_lowercase, token_probs=_lowercase, student=_lowercase, teacher=_lowercase )
distiller.train()
logger.info("""Let's go get some drinks.""" )
if __name__ == "__main__":
main()
| 66 | 0 |
import flax.linen as nn
import jax
import jax.numpy as jnp
class lowercase__( nn.Module ):
"""simple docstring"""
a :int
a :jnp.dtype = jnp.floataa
def _lowercase ( self : List[Any] ) -> Union[str, Any]:
lowercase_ = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : str , SCREAMING_SNAKE_CASE_ : Optional[Any] ) -> Optional[int]:
lowercase_ , lowercase_ , lowercase_ , lowercase_ = hidden_states.shape
lowercase_ = jax.image.resize(
SCREAMING_SNAKE_CASE_ , shape=(batch, height * 2, width * 2, channels) , method='''nearest''' , )
lowercase_ = self.conv(SCREAMING_SNAKE_CASE_ )
return hidden_states
class lowercase__( nn.Module ):
"""simple docstring"""
a :int
a :jnp.dtype = jnp.floataa
def _lowercase ( self : Union[str, Any] ) -> Union[str, Any]:
lowercase_ = nn.Conv(
self.out_channels , kernel_size=(3, 3) , strides=(2, 2) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
def __call__( self : Tuple , SCREAMING_SNAKE_CASE_ : Dict ) -> List[str]:
# pad = ((0, 0), (0, 1), (0, 1), (0, 0)) # pad height and width dim
# hidden_states = jnp.pad(hidden_states, pad_width=pad)
lowercase_ = self.conv(SCREAMING_SNAKE_CASE_ )
return hidden_states
class lowercase__( nn.Module ):
"""simple docstring"""
a :int
a :int = None
a :float = 0.0
a :bool = None
a :jnp.dtype = jnp.floataa
def _lowercase ( self : Any ) -> List[str]:
lowercase_ = self.in_channels if self.out_channels is None else self.out_channels
lowercase_ = nn.GroupNorm(num_groups=3_2 , epsilon=1e-5 )
lowercase_ = nn.Conv(
SCREAMING_SNAKE_CASE_ , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
lowercase_ = nn.Dense(SCREAMING_SNAKE_CASE_ , dtype=self.dtype )
lowercase_ = nn.GroupNorm(num_groups=3_2 , epsilon=1e-5 )
lowercase_ = nn.Dropout(self.dropout_prob )
lowercase_ = nn.Conv(
SCREAMING_SNAKE_CASE_ , kernel_size=(3, 3) , strides=(1, 1) , padding=((1, 1), (1, 1)) , dtype=self.dtype , )
lowercase_ = self.in_channels != out_channels if self.use_nin_shortcut is None else self.use_nin_shortcut
lowercase_ = None
if use_nin_shortcut:
lowercase_ = nn.Conv(
SCREAMING_SNAKE_CASE_ , kernel_size=(1, 1) , strides=(1, 1) , padding='''VALID''' , dtype=self.dtype , )
def __call__( self : Union[str, Any] , SCREAMING_SNAKE_CASE_ : Dict , SCREAMING_SNAKE_CASE_ : Any , SCREAMING_SNAKE_CASE_ : Optional[int]=True ) -> Tuple:
lowercase_ = hidden_states
lowercase_ = self.norma(SCREAMING_SNAKE_CASE_ )
lowercase_ = nn.swish(SCREAMING_SNAKE_CASE_ )
lowercase_ = self.conva(SCREAMING_SNAKE_CASE_ )
lowercase_ = self.time_emb_proj(nn.swish(SCREAMING_SNAKE_CASE_ ) )
lowercase_ = jnp.expand_dims(jnp.expand_dims(SCREAMING_SNAKE_CASE_ , 1 ) , 1 )
lowercase_ = hidden_states + temb
lowercase_ = self.norma(SCREAMING_SNAKE_CASE_ )
lowercase_ = nn.swish(SCREAMING_SNAKE_CASE_ )
lowercase_ = self.dropout(SCREAMING_SNAKE_CASE_ , SCREAMING_SNAKE_CASE_ )
lowercase_ = self.conva(SCREAMING_SNAKE_CASE_ )
if self.conv_shortcut is not None:
lowercase_ = self.conv_shortcut(SCREAMING_SNAKE_CASE_ )
return hidden_states + residual
| 30 |
"""simple docstring"""
import json
import os
import subprocess
import unittest
from ast import literal_eval
import pytest
from parameterized import parameterized, parameterized_class
from . import is_sagemaker_available
if is_sagemaker_available():
from sagemaker import Session, TrainingJobAnalytics
from sagemaker.huggingface import HuggingFace
@pytest.mark.skipif(
literal_eval(os.getenv("""TEST_SAGEMAKER""" , """False""" ) ) is not True , reason="""Skipping test because should only be run when releasing minor transformers version""" , )
@pytest.mark.usefixtures("""sm_env""" )
@parameterized_class(
[
{
"""framework""": """pytorch""",
"""script""": """run_glue_model_parallelism.py""",
"""model_name_or_path""": """roberta-large""",
"""instance_type""": """ml.p3dn.24xlarge""",
"""results""": {"""train_runtime""": 1_6_0_0, """eval_accuracy""": 0.3, """eval_loss""": 1.2},
},
{
"""framework""": """pytorch""",
"""script""": """run_glue.py""",
"""model_name_or_path""": """roberta-large""",
"""instance_type""": """ml.p3dn.24xlarge""",
"""results""": {"""train_runtime""": 1_6_0_0, """eval_accuracy""": 0.3, """eval_loss""": 1.2},
},
] )
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: Any ) -> str:
if self.framework == "pytorch":
subprocess.run(
f"""cp ./examples/pytorch/text-classification/run_glue.py {self.env.test_path}/run_glue.py""".split() , encoding="""utf-8""" , check=snake_case , )
assert hasattr(self , """env""" )
def lowerCAmelCase_ ( self: int , snake_case: Dict ) -> List[Any]:
# configuration for running training on smdistributed Model Parallel
snake_case_ :Tuple = {
"""enabled""": True,
"""processes_per_host""": 8,
}
snake_case_ :List[Any] = {
"""enabled""": True,
"""parameters""": {
"""microbatches""": 4,
"""placement_strategy""": """spread""",
"""pipeline""": """interleaved""",
"""optimize""": """speed""",
"""partitions""": 4,
"""ddp""": True,
},
}
snake_case_ :Tuple = {"""smdistributed""": {"""modelparallel""": smp_options}, """mpi""": mpi_options}
snake_case_ :Any = """trainer""" if self.script == """run_glue.py""" else """smtrainer"""
# creates estimator
return HuggingFace(
entry_point=self.script , source_dir=self.env.test_path , role=self.env.role , image_uri=self.env.image_uri , base_job_name=f"""{self.env.base_job_name}-{instance_count}-smp-{name_extension}""" , instance_count=snake_case , instance_type=self.instance_type , debugger_hook_config=snake_case , hyperparameters={
**self.env.hyperparameters,
"""model_name_or_path""": self.model_name_or_path,
"""max_steps""": 500,
} , metric_definitions=self.env.metric_definitions , distribution=snake_case , py_version="""py36""" , )
def lowerCAmelCase_ ( self: Any , snake_case: Tuple ) -> List[str]:
TrainingJobAnalytics(snake_case ).export_csv(f"""{self.env.test_path}/{job_name}_metrics.csv""" )
@parameterized.expand([(1,)] )
def lowerCAmelCase_ ( self: Dict , snake_case: Dict ) -> List[Any]:
# create estimator
snake_case_ :List[Any] = self.create_estimator(snake_case )
# run training
estimator.fit()
# result dataframe
snake_case_ :Any = TrainingJobAnalytics(estimator.latest_training_job.name ).dataframe()
# extract kpis
snake_case_ :Tuple = list(result_metrics_df[result_metrics_df.metric_name == """eval_accuracy"""]["""value"""] )
snake_case_ :Dict = list(result_metrics_df[result_metrics_df.metric_name == """eval_loss"""]["""value"""] )
# get train time from SageMaker job, this includes starting, preprocessing, stopping
snake_case_ :int = (
Session().describe_training_job(estimator.latest_training_job.name ).get("""TrainingTimeInSeconds""" , 999_999 )
)
# assert kpis
assert train_runtime <= self.results["train_runtime"]
assert all(t >= self.results["""eval_accuracy"""] for t in eval_accuracy )
assert all(t <= self.results["""eval_loss"""] for t in eval_loss )
# dump tests result into json file to share in PR
with open(f"""{estimator.latest_training_job.name}.json""" , """w""" ) as outfile:
json.dump({"""train_time""": train_runtime, """eval_accuracy""": eval_accuracy, """eval_loss""": eval_loss} , snake_case )
| 66 | 0 |
'''simple docstring'''
import torch
from transformers import PreTrainedModel, XLMRobertaConfig, XLMRobertaModel
class lowerCamelCase_ (snake_case__ ):
'''simple docstring'''
__UpperCamelCase: Tuple = "M-CLIP"
def __init__( self : Optional[Any] , A : List[Any]=1024 , A : Any=768 , **A : Tuple ):
_UpperCAmelCase : str = transformerDimSize
_UpperCAmelCase : Optional[int] = imageDimSize
super().__init__(**A )
class lowerCamelCase_ (snake_case__ ):
'''simple docstring'''
__UpperCamelCase: Dict = MCLIPConfig
def __init__( self : Optional[Any] , A : Any , *A : Any , **A : Optional[int] ):
super().__init__(A , *A , **A )
_UpperCAmelCase : Union[str, Any] = XLMRobertaModel(A )
_UpperCAmelCase : Optional[int] = torch.nn.Linear(
in_features=config.transformerDimensions , out_features=config.numDims )
def _A ( self : Union[str, Any] , A : int , A : Any ):
_UpperCAmelCase : Tuple = self.transformer(input_ids=A , attention_mask=A )[0]
_UpperCAmelCase : List[str] = (embs * attention_mask.unsqueeze(2 )).sum(dim=1 ) / attention_mask.sum(dim=1 )[:, None]
return self.LinearTransformation(A ), embs
| 31 |
"""simple docstring"""
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 lowerCamelCase :
'''simple docstring'''
def __init__( self: Dict , snake_case: Optional[Any] , snake_case: Tuple=13 , snake_case: Any=32 , snake_case: Union[str, Any]=2 , snake_case: Tuple=3 , snake_case: Union[str, Any]=16 , snake_case: Union[str, Any]=[1, 2, 1] , snake_case: Optional[Any]=[2, 2, 4] , snake_case: str=2 , snake_case: List[str]=2.0 , snake_case: Optional[int]=True , snake_case: Union[str, Any]=0.0 , snake_case: Optional[int]=0.0 , snake_case: Optional[Any]=0.1 , snake_case: List[str]="gelu" , snake_case: Any=False , snake_case: Optional[Any]=True , snake_case: Optional[int]=0.0_2 , snake_case: Any=1E-5 , snake_case: Optional[int]=True , snake_case: int=None , snake_case: Any=True , snake_case: str=10 , snake_case: Optional[Any]=8 , snake_case: Union[str, Any]=["stage1", "stage2", "stage3"] , snake_case: Tuple=[1, 2, 3] , ) -> Dict:
snake_case_ :Dict = parent
snake_case_ :List[Any] = batch_size
snake_case_ :Dict = image_size
snake_case_ :Dict = patch_size
snake_case_ :Tuple = num_channels
snake_case_ :List[Any] = embed_dim
snake_case_ :List[str] = depths
snake_case_ :str = num_heads
snake_case_ :Tuple = window_size
snake_case_ :Tuple = mlp_ratio
snake_case_ :int = qkv_bias
snake_case_ :Tuple = hidden_dropout_prob
snake_case_ :Optional[Any] = attention_probs_dropout_prob
snake_case_ :Dict = drop_path_rate
snake_case_ :Any = hidden_act
snake_case_ :Any = use_absolute_embeddings
snake_case_ :int = patch_norm
snake_case_ :List[Any] = layer_norm_eps
snake_case_ :Tuple = initializer_range
snake_case_ :str = is_training
snake_case_ :int = scope
snake_case_ :Tuple = use_labels
snake_case_ :Tuple = type_sequence_label_size
snake_case_ :str = encoder_stride
snake_case_ :List[Any] = out_features
snake_case_ :str = out_indices
def lowerCAmelCase_ ( self: Tuple ) -> Dict:
snake_case_ :Optional[int] = floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size] )
snake_case_ :str = None
if self.use_labels:
snake_case_ :Optional[int] = ids_tensor([self.batch_size] , self.type_sequence_label_size )
snake_case_ :Union[str, Any] = self.get_config()
return config, pixel_values, labels
def lowerCAmelCase_ ( self: int ) -> Optional[Any]:
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 lowerCAmelCase_ ( self: List[Any] , snake_case: str , snake_case: int , snake_case: List[str] ) -> Any:
snake_case_ :Dict = MaskFormerSwinModel(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Tuple = model(snake_case )
snake_case_ :Dict = ((config.image_size // config.patch_size) ** 2) // (4 ** (len(config.depths ) - 1))
snake_case_ :Any = 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 lowerCAmelCase_ ( self: Optional[Any] , snake_case: int , snake_case: List[str] , snake_case: Tuple ) -> Union[str, Any]:
snake_case_ :Any = MaskFormerSwinBackbone(config=snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Optional[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_ :Optional[Any] = ["""stem"""]
snake_case_ :str = MaskFormerSwinBackbone(config=snake_case )
def lowerCAmelCase_ ( self: List[str] ) -> Optional[Any]:
snake_case_ :Optional[int] = self.prepare_config_and_inputs()
snake_case_, snake_case_, snake_case_ :str = config_and_inputs
snake_case_ :Tuple = {"""pixel_values""": pixel_values}
return config, inputs_dict
@require_torch
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : Union[str, Any] = (
(
MaskFormerSwinModel,
MaskFormerSwinBackbone,
)
if is_torch_available()
else ()
)
_A : str = {"""feature-extraction""": MaskFormerSwinModel} if is_torch_available() else {}
_A : List[str] = False
_A : Any = False
_A : Dict = False
_A : List[Any] = False
_A : Optional[int] = False
def lowerCAmelCase_ ( self: Dict ) -> Any:
snake_case_ :str = MaskFormerSwinModelTester(self )
snake_case_ :Optional[Any] = 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 lowerCAmelCase_ ( self: List[str] ) -> Optional[int]:
pass
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Dict:
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 lowerCAmelCase_ ( self: Any ) -> Tuple:
return
def lowerCAmelCase_ ( self: Any ) -> Any:
snake_case_ :List[Any] = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*snake_case )
def lowerCAmelCase_ ( self: Union[str, Any] ) -> int:
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 lowerCAmelCase_ ( self: str ) -> List[str]:
pass
@unittest.skip("""Swin does not support feedforward chunking""" )
def lowerCAmelCase_ ( self: int ) -> Optional[int]:
pass
def lowerCAmelCase_ ( self: List[str] ) -> List[Any]:
snake_case_, snake_case_ :List[Any] = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :str = model_class(snake_case )
self.assertIsInstance(model.get_input_embeddings() , (nn.Module) )
snake_case_ :Dict = model.get_output_embeddings()
self.assertTrue(x is None or isinstance(snake_case , nn.Linear ) )
def lowerCAmelCase_ ( self: Tuple ) -> Dict:
snake_case_, snake_case_ :int = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
snake_case_ :Optional[int] = model_class(snake_case )
snake_case_ :str = inspect.signature(model.forward )
# signature.parameters is an OrderedDict => so arg_names order is deterministic
snake_case_ :str = [*signature.parameters.keys()]
snake_case_ :str = ["""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 lowerCAmelCase_ ( self: List[Any] ) -> List[Any]:
pass
@unittest.skip(reason="""MaskFormerSwin is only used as an internal backbone""" )
def lowerCAmelCase_ ( self: Dict ) -> List[Any]:
pass
def lowerCAmelCase_ ( self: Union[str, Any] , snake_case: Union[str, Any] , snake_case: int , snake_case: Any , snake_case: List[str] ) -> str:
snake_case_ :List[str] = model_class(snake_case )
model.to(snake_case )
model.eval()
with torch.no_grad():
snake_case_ :List[Any] = model(**self._prepare_for_class(snake_case , snake_case ) )
snake_case_ :Any = outputs.hidden_states
snake_case_ :Optional[int] = 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_ :str = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :int = (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 lowerCAmelCase_ ( self: List[Any] ) -> Optional[int]:
snake_case_, snake_case_ :Any = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :List[Any] = (
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_ :Tuple = 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_ :List[Any] = True
self.check_hidden_states_output(snake_case , snake_case , snake_case , snake_case )
def lowerCAmelCase_ ( self: Optional[Any] ) -> Tuple:
snake_case_, snake_case_ :int = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :List[Any] = 3
snake_case_ :List[Any] = (
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_ :Any = (
config.patch_size
if isinstance(config.patch_size , collections.abc.Iterable )
else (config.patch_size, config.patch_size)
)
snake_case_ :Tuple = image_size[0] + patch_size[0] - (image_size[0] % patch_size[0])
snake_case_ :List[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_ :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 lowerCAmelCase_ ( self: Union[str, Any] ) -> List[str]:
pass
@unittest.skip(reason="""This will be fixed once MaskFormerSwin is replaced by native Swin""" )
def lowerCAmelCase_ ( self: List[str] ) -> str:
pass
@unittest.skip(reason="""This will be fixed once MaskFormerSwin is replaced by native Swin""" )
def lowerCAmelCase_ ( self: str ) -> List[Any]:
pass
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Optional[Any]:
snake_case_, snake_case_ :Dict = self.model_tester.prepare_config_and_inputs_for_common()
def set_nan_tensor_to_zero(snake_case: str ):
snake_case_ :Optional[int] = 0
return t
def check_equivalence(snake_case: List[Any] , snake_case: Union[str, Any] , snake_case: int , snake_case: Tuple={} ):
with torch.no_grad():
snake_case_ :List[Any] = model(**snake_case , return_dict=snake_case , **snake_case )
snake_case_ :Any = model(**snake_case , return_dict=snake_case , **snake_case ).to_tuple()
def recursive_check(snake_case: List[Any] , snake_case: int ):
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_ :int = model_class(snake_case )
model.to(snake_case )
model.eval()
snake_case_ :Any = 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 )
snake_case_ :Tuple = self._prepare_for_class(snake_case , snake_case , return_labels=snake_case )
snake_case_ :Dict = self._prepare_for_class(snake_case , snake_case , return_labels=snake_case )
check_equivalence(snake_case , snake_case , snake_case )
snake_case_ :Tuple = self._prepare_for_class(snake_case , snake_case )
snake_case_ :Any = self._prepare_for_class(snake_case , snake_case )
check_equivalence(snake_case , snake_case , snake_case , {"""output_hidden_states""": True} )
snake_case_ :Dict = 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 lowerCamelCase ( unittest.TestCase , _lowerCAmelCase ):
'''simple docstring'''
_A : int = (MaskFormerSwinBackbone,) if is_torch_available() else ()
_A : Tuple = MaskFormerSwinConfig
def lowerCAmelCase_ ( self: List[str] ) -> Optional[int]:
snake_case_ :Optional[Any] = MaskFormerSwinModelTester(self )
def lowerCAmelCase_ ( self: int ) -> Optional[int]:
snake_case_, snake_case_ :Any = self.model_tester.prepare_config_and_inputs_for_common()
snake_case_ :Tuple = inputs_dict["""pixel_values"""].shape[0]
for backbone_class in self.all_model_classes:
snake_case_ :List[str] = backbone_class(snake_case )
backbone.to(snake_case )
backbone.eval()
snake_case_ :List[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_ :Union[str, Any] = 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_ :List[Any] = 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_ :List[Any] = backbone(**snake_case , output_attentions=snake_case )
self.assertIsNotNone(outputs.attentions )
| 66 | 0 |
import os
import pickle
import unittest
from transformers import AutoTokenizer
from transformers.models.bert.tokenization_bert import BertTokenizer
from transformers.models.bert_japanese.tokenization_bert_japanese import (
VOCAB_FILES_NAMES,
BertJapaneseTokenizer,
CharacterTokenizer,
JumanppTokenizer,
MecabTokenizer,
SudachiTokenizer,
WordpieceTokenizer,
)
from transformers.testing_utils import custom_tokenizers, require_jumanpp, require_sudachi
from ...test_tokenization_common import TokenizerTesterMixin
@custom_tokenizers
class SCREAMING_SNAKE_CASE__ ( lowercase__ , unittest.TestCase ):
snake_case__ : int = BertJapaneseTokenizer
snake_case__ : Optional[int] = False
snake_case__ : Union[str, Any] = True
def SCREAMING_SNAKE_CASE ( self : Tuple ) -> List[str]:
super().setUp()
a_ : Optional[Any] = [
'[UNK]',
'[CLS]',
'[SEP]',
'こんにちは',
'こん',
'にちは',
'ばんは',
'##こん',
'##にちは',
'##ばんは',
'世界',
'##世界',
'、',
'##、',
'。',
'##。',
]
a_ : Dict = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['vocab_file'] )
with open(self.vocab_file , 'w' , encoding='utf-8' ) as vocab_writer:
vocab_writer.write(''.join([x + '\n' for x in vocab_tokens] ) )
def SCREAMING_SNAKE_CASE ( self : str , SCREAMING_SNAKE_CASE__ : List[Any] ) -> Optional[int]:
a_ : Optional[int] = 'こんにちは、世界。 \nこんばんは、世界。'
a_ : Optional[Any] = 'こんにちは 、 世界 。 こんばんは 、 世界 。'
return input_text, output_text
def SCREAMING_SNAKE_CASE ( self : Dict , SCREAMING_SNAKE_CASE__ : List[str] ) -> int:
a_ , a_ : int = self.get_input_output_texts(SCREAMING_SNAKE_CASE__ )
a_ : List[Any] = tokenizer.encode(SCREAMING_SNAKE_CASE__ , add_special_tokens=SCREAMING_SNAKE_CASE__ )
a_ : Tuple = tokenizer.decode(SCREAMING_SNAKE_CASE__ , clean_up_tokenization_spaces=SCREAMING_SNAKE_CASE__ )
return text, ids
def SCREAMING_SNAKE_CASE ( self : Optional[int] ) -> Dict:
pass # TODO add if relevant
def SCREAMING_SNAKE_CASE ( self : List[str] ) -> Tuple:
pass # TODO add if relevant
def SCREAMING_SNAKE_CASE ( self : List[Any] ) -> Optional[Any]:
pass # TODO add if relevant
def SCREAMING_SNAKE_CASE ( self : Any ) -> Any:
a_ : Union[str, Any] = self.tokenizer_class(self.vocab_file )
a_ : str = tokenizer.tokenize('こんにちは、世界。\nこんばんは、世界。' )
self.assertListEqual(SCREAMING_SNAKE_CASE__ , ['こんにちは', '、', '世界', '。', 'こん', '##ばんは', '、', '世界', '。'] )
self.assertListEqual(tokenizer.convert_tokens_to_ids(SCREAMING_SNAKE_CASE__ ) , [3, 1_2, 1_0, 1_4, 4, 9, 1_2, 1_0, 1_4] )
def SCREAMING_SNAKE_CASE ( self : Optional[int] ) -> List[str]:
a_ : str = self.tokenizer_class(self.vocab_file , word_tokenizer_type='mecab' )
self.assertIsNotNone(SCREAMING_SNAKE_CASE__ )
a_ : str = 'こんにちは、世界。\nこんばんは、世界。'
a_ : List[Any] = tokenizer.tokenize(SCREAMING_SNAKE_CASE__ )
self.assertListEqual(SCREAMING_SNAKE_CASE__ , ['こんにちは', '、', '世界', '。', 'こん', '##ばんは', '、', '世界', '。'] )
self.assertListEqual(tokenizer.convert_tokens_to_ids(SCREAMING_SNAKE_CASE__ ) , [3, 1_2, 1_0, 1_4, 4, 9, 1_2, 1_0, 1_4] )
a_ : Optional[int] = os.path.join(self.tmpdirname , 'tokenizer.bin' )
with open(SCREAMING_SNAKE_CASE__ , 'wb' ) as handle:
pickle.dump(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ )
with open(SCREAMING_SNAKE_CASE__ , 'rb' ) as handle:
a_ : Union[str, Any] = pickle.load(SCREAMING_SNAKE_CASE__ )
a_ : List[Any] = tokenizer_new.tokenize(SCREAMING_SNAKE_CASE__ )
self.assertListEqual(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ )
def SCREAMING_SNAKE_CASE ( self : Tuple ) -> Union[str, Any]:
a_ : List[str] = MecabTokenizer(mecab_dic='ipadic' )
self.assertListEqual(
tokenizer.tokenize(' \tｱｯﾌﾟﾙストアでiPhone８ が \n 発売された 。 ' ) , ['アップルストア', 'で', 'iPhone', '8', 'が', '発売', 'さ', 'れ', 'た', '。'] , )
def SCREAMING_SNAKE_CASE ( self : Union[str, Any] ) -> List[Any]:
try:
a_ : List[str] = MecabTokenizer(mecab_dic='unidic_lite' )
except ModuleNotFoundError:
return
self.assertListEqual(
tokenizer.tokenize(' \tｱｯﾌﾟﾙストアでiPhone８ が \n 発売された 。 ' ) , ['アップル', 'ストア', 'で', 'iPhone', '8', 'が', '発売', 'さ', 'れ', 'た', '。'] , )
def SCREAMING_SNAKE_CASE ( self : Tuple ) -> Optional[int]:
try:
a_ : Union[str, Any] = MecabTokenizer(mecab_dic='unidic' )
except ModuleNotFoundError:
return
self.assertListEqual(
tokenizer.tokenize(' \tｱｯﾌﾟﾙストアでiPhone８ が \n 発売された 。 ' ) , ['アップル', 'ストア', 'で', 'iPhone', '8', 'が', '発売', 'さ', 'れ', 'た', '。'] , )
def SCREAMING_SNAKE_CASE ( self : str ) -> Any:
a_ : Optional[Any] = MecabTokenizer(do_lower_case=SCREAMING_SNAKE_CASE__ , mecab_dic='ipadic' )
self.assertListEqual(
tokenizer.tokenize(' \tｱｯﾌﾟﾙストアでiPhone８ が \n 発売された 。 ' ) , ['アップルストア', 'で', 'iphone', '8', 'が', '発売', 'さ', 'れ', 'た', '。'] , )
def SCREAMING_SNAKE_CASE ( self : Optional[int] ) -> Union[str, Any]:
try:
a_ : Any = MecabTokenizer(
do_lower_case=SCREAMING_SNAKE_CASE__ , normalize_text=SCREAMING_SNAKE_CASE__ , mecab_option='-d /usr/local/lib/mecab/dic/jumandic' )
except RuntimeError:
# if dict doesn't exist in the system, previous code raises this error.
return
self.assertListEqual(
tokenizer.tokenize(' \tｱｯﾌﾟﾙストアでiPhone８ が \n 発売された 。 ' ) , ['ｱｯﾌﾟﾙストア', 'で', 'iPhone', '８', 'が', '発売', 'さ', 'れた', '\u3000', '。'] , )
def SCREAMING_SNAKE_CASE ( self : List[str] ) -> Any:
a_ : int = MecabTokenizer(normalize_text=SCREAMING_SNAKE_CASE__ , mecab_dic='ipadic' )
self.assertListEqual(
tokenizer.tokenize(' \tｱｯﾌﾟﾙストアでiPhone８ が \n 発売された 。 ' ) , ['ｱｯﾌﾟﾙストア', 'で', 'iPhone', '８', 'が', '発売', 'さ', 'れ', 'た', ' ', '。'] , )
@require_sudachi
def SCREAMING_SNAKE_CASE ( self : Optional[int] ) -> str:
a_ : List[str] = self.tokenizer_class(self.vocab_file , word_tokenizer_type='sudachi' )
self.assertIsNotNone(SCREAMING_SNAKE_CASE__ )
a_ : List[str] = 'こんにちは、世界。\nこんばんは、世界。'
a_ : Dict = tokenizer.tokenize(SCREAMING_SNAKE_CASE__ )
self.assertListEqual(SCREAMING_SNAKE_CASE__ , ['こんにちは', '、', '世界', '。', 'こん', '##ばんは', '、', '世界', '。'] )
self.assertListEqual(tokenizer.convert_tokens_to_ids(SCREAMING_SNAKE_CASE__ ) , [3, 1_2, 1_0, 1_4, 4, 9, 1_2, 1_0, 1_4] )
a_ : int = os.path.join(self.tmpdirname , 'tokenizer.bin' )
with open(SCREAMING_SNAKE_CASE__ , 'wb' ) as handle:
pickle.dump(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ )
with open(SCREAMING_SNAKE_CASE__ , 'rb' ) as handle:
a_ : Union[str, Any] = pickle.load(SCREAMING_SNAKE_CASE__ )
a_ : Optional[Any] = tokenizer_new.tokenize(SCREAMING_SNAKE_CASE__ )
self.assertListEqual(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ )
@require_sudachi
def SCREAMING_SNAKE_CASE ( self : int ) -> List[str]:
a_ : List[str] = SudachiTokenizer(sudachi_dict_type='core' )
self.assertListEqual(
tokenizer.tokenize(' \tｱｯﾌﾟﾙストアでiPhone８ が \n 発売された 。 ' ) , [' ', '\t', 'アップル', 'ストア', 'で', 'iPhone', '8', ' ', 'が', ' ', ' ', '\n ', '発売', 'さ', 'れ', 'た', ' ', '。', ' ', ' '] , )
@require_sudachi
def SCREAMING_SNAKE_CASE ( self : Tuple ) -> Union[str, Any]:
a_ : List[Any] = SudachiTokenizer(sudachi_dict_type='core' , sudachi_split_mode='A' )
self.assertListEqual(tokenizer.tokenize('外国人参政権' ) , ['外国', '人', '参政', '権'] )
@require_sudachi
def SCREAMING_SNAKE_CASE ( self : Union[str, Any] ) -> Tuple:
a_ : Tuple = SudachiTokenizer(sudachi_dict_type='core' , sudachi_split_mode='B' )
self.assertListEqual(tokenizer.tokenize('外国人参政権' ) , ['外国人', '参政権'] )
@require_sudachi
def SCREAMING_SNAKE_CASE ( self : str ) -> List[str]:
a_ : Union[str, Any] = SudachiTokenizer(sudachi_dict_type='core' , sudachi_split_mode='C' )
self.assertListEqual(tokenizer.tokenize('外国人参政権' ) , ['外国人参政権'] )
@require_sudachi
def SCREAMING_SNAKE_CASE ( self : List[Any] ) -> Union[str, Any]:
a_ : Dict = SudachiTokenizer(do_lower_case=SCREAMING_SNAKE_CASE__ , sudachi_dict_type='core' )
self.assertListEqual(
tokenizer.tokenize(' \tｱｯﾌﾟﾙストアでiPhone８ が \n 発売された 。 ' ) , [' ', '\t', 'アップル', 'ストア', 'で', 'iphone', '8', ' ', 'が', ' ', ' ', '\n ', '発売', 'さ', 'れ', 'た', ' ', '。', ' ', ' '] , )
@require_sudachi
def SCREAMING_SNAKE_CASE ( self : Union[str, Any] ) -> Optional[Any]:
a_ : Any = SudachiTokenizer(normalize_text=SCREAMING_SNAKE_CASE__ , sudachi_dict_type='core' )
self.assertListEqual(
tokenizer.tokenize(' \tｱｯﾌﾟﾙストアでiPhone８ が \n 発売された 。 ' ) , [' ', '\t', 'ｱｯﾌﾟﾙ', 'ストア', 'で', 'iPhone', '８', ' ', 'が', ' ', ' ', '\n ', '発売', 'さ', 'れ', 'た', '\u3000', '。', ' ', ' '] , )
@require_sudachi
def SCREAMING_SNAKE_CASE ( self : Tuple ) -> List[str]:
a_ : int = SudachiTokenizer(trim_whitespace=SCREAMING_SNAKE_CASE__ , sudachi_dict_type='core' )
self.assertListEqual(
tokenizer.tokenize(' \tｱｯﾌﾟﾙストアでiPhone８ が \n 発売された 。 ' ) , ['アップル', 'ストア', 'で', 'iPhone', '8', 'が', '発売', 'さ', 'れ', 'た', '。'] , )
@require_jumanpp
def SCREAMING_SNAKE_CASE ( self : Tuple ) -> Optional[Any]:
a_ : Tuple = self.tokenizer_class(self.vocab_file , word_tokenizer_type='jumanpp' )
self.assertIsNotNone(SCREAMING_SNAKE_CASE__ )
a_ : Any = 'こんにちは、世界。\nこんばんは、世界。'
a_ : List[Any] = tokenizer.tokenize(SCREAMING_SNAKE_CASE__ )
self.assertListEqual(SCREAMING_SNAKE_CASE__ , ['こんにちは', '、', '世界', '。', 'こん', '##ばんは', '、', '世界', '。'] )
self.assertListEqual(tokenizer.convert_tokens_to_ids(SCREAMING_SNAKE_CASE__ ) , [3, 1_2, 1_0, 1_4, 4, 9, 1_2, 1_0, 1_4] )
a_ : Optional[int] = os.path.join(self.tmpdirname , 'tokenizer.bin' )
with open(SCREAMING_SNAKE_CASE__ , 'wb' ) as handle:
pickle.dump(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ )
with open(SCREAMING_SNAKE_CASE__ , 'rb' ) as handle:
a_ : Optional[Any] = pickle.load(SCREAMING_SNAKE_CASE__ )
a_ : Tuple = tokenizer_new.tokenize(SCREAMING_SNAKE_CASE__ )
self.assertListEqual(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ )
@require_jumanpp
def SCREAMING_SNAKE_CASE ( self : Union[str, Any] ) -> Optional[int]:
a_ : Optional[Any] = JumanppTokenizer()
self.assertListEqual(
tokenizer.tokenize(' \tｱｯﾌﾟﾙストアでiPhone８ が \n 発売された 。 ' ) , ['アップル', 'ストア', 'で', 'iPhone', '8', '\u3000', 'が', '\u3000', '\u3000', '\u3000', '発売', 'さ', 'れた', '\u3000', '。'] , )
@require_jumanpp
def SCREAMING_SNAKE_CASE ( self : Any ) -> Optional[Any]:
a_ : Union[str, Any] = JumanppTokenizer(do_lower_case=SCREAMING_SNAKE_CASE__ )
self.assertListEqual(
tokenizer.tokenize(' \tｱｯﾌﾟﾙストアでiPhone８ が \n 発売された 。 ' ) , ['アップル', 'ストア', 'で', 'iphone', '8', '\u3000', 'が', '\u3000', '\u3000', '\u3000', '発売', 'さ', 'れた', '\u3000', '。'] , )
@require_jumanpp
def SCREAMING_SNAKE_CASE ( self : List[Any] ) -> List[Any]:
a_ : Optional[int] = JumanppTokenizer(normalize_text=SCREAMING_SNAKE_CASE__ )
self.assertListEqual(
tokenizer.tokenize(' \tｱｯﾌﾟﾙストアでiPhone８ が \n 発売された 。 ' ) , ['ｱ', 'ｯ', 'ﾌ', 'ﾟ', 'ﾙ', 'ストア', 'で', 'iPhone', '８', '\u3000', 'が', '\u3000', '\u3000', '\u3000', '発売', 'さ', 'れた', '\u3000', '。'] , )
@require_jumanpp
def SCREAMING_SNAKE_CASE ( self : Any ) -> int:
a_ : Dict = JumanppTokenizer(trim_whitespace=SCREAMING_SNAKE_CASE__ )
self.assertListEqual(
tokenizer.tokenize(' \tｱｯﾌﾟﾙストアでiPhone８ が \n 発売された 。 ' ) , ['アップル', 'ストア', 'で', 'iPhone', '8', 'が', '発売', 'さ', 'れた', '。'] , )
@require_jumanpp
def SCREAMING_SNAKE_CASE ( self : Any ) -> List[str]:
a_ : str = JumanppTokenizer()
self.assertListEqual(
tokenizer.tokenize('ありがとうございますm(_ _)ｍ見つけるのが大変です。' ) , ['ありがとう', 'ございます', 'm(_ _)m', '見つける', 'の', 'が', '大変です', '。'] , )
def SCREAMING_SNAKE_CASE ( self : Dict ) -> Any:
a_ : Tuple = ['[UNK]', '[CLS]', '[SEP]', 'こんにちは', 'こん', 'にちは', 'ばんは', '##こん', '##にちは', '##ばんは']
a_ : List[Any] = {}
for i, token in enumerate(SCREAMING_SNAKE_CASE__ ):
a_ : List[str] = i
a_ : Dict = WordpieceTokenizer(vocab=SCREAMING_SNAKE_CASE__ , unk_token='[UNK]' )
self.assertListEqual(tokenizer.tokenize('' ) , [] )
self.assertListEqual(tokenizer.tokenize('こんにちは' ) , ['こんにちは'] )
self.assertListEqual(tokenizer.tokenize('こんばんは' ) , ['こん', '##ばんは'] )
self.assertListEqual(tokenizer.tokenize('こんばんは こんばんにちは こんにちは' ) , ['こん', '##ばんは', '[UNK]', 'こんにちは'] )
def SCREAMING_SNAKE_CASE ( self : Union[str, Any] ) -> List[str]:
a_ : List[str] = BertJapaneseTokenizer.from_pretrained('nlp-waseda/roberta-base-japanese-with-auto-jumanpp' )
a_ : List[str] = tokenizer.subword_tokenizer
a_ : Optional[int] = subword_tokenizer.tokenize('国境 の 長い トンネル を 抜ける と 雪国 であった 。' )
self.assertListEqual(SCREAMING_SNAKE_CASE__ , ['▁国境', '▁の', '▁長い', '▁トンネル', '▁を', '▁抜ける', '▁と', '▁雪', '国', '▁であった', '▁。'] )
a_ : Optional[int] = subword_tokenizer.tokenize('こんばんは こんばん にち は こんにちは' )
self.assertListEqual(SCREAMING_SNAKE_CASE__ , ['▁こん', 'ばん', 'は', '▁こん', 'ばん', '▁に', 'ち', '▁は', '▁こんにちは'] )
def SCREAMING_SNAKE_CASE ( self : Tuple ) -> Tuple:
a_ : Optional[Any] = self.tokenizer_class.from_pretrained('cl-tohoku/bert-base-japanese' )
a_ : int = tokenizer.encode('ありがとう。' , add_special_tokens=SCREAMING_SNAKE_CASE__ )
a_ : Dict = tokenizer.encode('どういたしまして。' , add_special_tokens=SCREAMING_SNAKE_CASE__ )
a_ : int = tokenizer.build_inputs_with_special_tokens(SCREAMING_SNAKE_CASE__ )
a_ : List[Any] = tokenizer.build_inputs_with_special_tokens(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ )
# 2 is for "[CLS]", 3 is for "[SEP]"
assert encoded_sentence == [2] + text + [3]
assert encoded_pair == [2] + text + [3] + text_a + [3]
@custom_tokenizers
class SCREAMING_SNAKE_CASE__ ( lowercase__ , unittest.TestCase ):
snake_case__ : Optional[int] = BertJapaneseTokenizer
snake_case__ : str = False
def SCREAMING_SNAKE_CASE ( self : str ) -> Any:
super().setUp()
a_ : int = ['[UNK]', '[CLS]', '[SEP]', 'こ', 'ん', 'に', 'ち', 'は', 'ば', '世', '界', '、', '。']
a_ : str = os.path.join(self.tmpdirname , VOCAB_FILES_NAMES['vocab_file'] )
with open(self.vocab_file , 'w' , encoding='utf-8' ) as vocab_writer:
vocab_writer.write(''.join([x + '\n' for x in vocab_tokens] ) )
def SCREAMING_SNAKE_CASE ( self : str , **SCREAMING_SNAKE_CASE__ : Optional[int] ) -> Tuple:
return BertJapaneseTokenizer.from_pretrained(self.tmpdirname , subword_tokenizer_type='character' , **SCREAMING_SNAKE_CASE__ )
def SCREAMING_SNAKE_CASE ( self : Dict , SCREAMING_SNAKE_CASE__ : List[str] ) -> List[str]:
a_ : List[Any] = 'こんにちは、世界。 \nこんばんは、世界。'
a_ : Optional[Any] = 'こ ん に ち は 、 世 界 。 こ ん ば ん は 、 世 界 。'
return input_text, output_text
def SCREAMING_SNAKE_CASE ( self : Any ) -> List[str]:
pass # TODO add if relevant
def SCREAMING_SNAKE_CASE ( self : List[Any] ) -> List[str]:
pass # TODO add if relevant
def SCREAMING_SNAKE_CASE ( self : Optional[Any] ) -> Tuple:
pass # TODO add if relevant
def SCREAMING_SNAKE_CASE ( self : Optional[Any] ) -> Dict:
a_ : Optional[int] = self.tokenizer_class(self.vocab_file , subword_tokenizer_type='character' )
a_ : Any = tokenizer.tokenize('こんにちは、世界。 \nこんばんは、世界。' )
self.assertListEqual(
SCREAMING_SNAKE_CASE__ , ['こ', 'ん', 'に', 'ち', 'は', '、', '世', '界', '。', 'こ', 'ん', 'ば', 'ん', 'は', '、', '世', '界', '。'] )
self.assertListEqual(
tokenizer.convert_tokens_to_ids(SCREAMING_SNAKE_CASE__ ) , [3, 4, 5, 6, 7, 1_1, 9, 1_0, 1_2, 3, 4, 8, 4, 7, 1_1, 9, 1_0, 1_2] )
def SCREAMING_SNAKE_CASE ( self : str ) -> Optional[Any]:
a_ : Dict = ['[UNK]', '[CLS]', '[SEP]', 'こ', 'ん', 'に', 'ち', 'は', 'ば', '世', '界', '、', '。']
a_ : Dict = {}
for i, token in enumerate(SCREAMING_SNAKE_CASE__ ):
a_ : str = i
a_ : Optional[Any] = CharacterTokenizer(vocab=SCREAMING_SNAKE_CASE__ , unk_token='[UNK]' )
self.assertListEqual(tokenizer.tokenize('' ) , [] )
self.assertListEqual(tokenizer.tokenize('こんにちは' ) , ['こ', 'ん', 'に', 'ち', 'は'] )
self.assertListEqual(tokenizer.tokenize('こんにちほ' ) , ['こ', 'ん', 'に', 'ち', '[UNK]'] )
def SCREAMING_SNAKE_CASE ( self : str ) -> Optional[Any]:
a_ : Union[str, Any] = self.tokenizer_class.from_pretrained('cl-tohoku/bert-base-japanese-char' )
a_ : Optional[Any] = tokenizer.encode('ありがとう。' , add_special_tokens=SCREAMING_SNAKE_CASE__ )
a_ : List[Any] = tokenizer.encode('どういたしまして。' , add_special_tokens=SCREAMING_SNAKE_CASE__ )
a_ : Union[str, Any] = tokenizer.build_inputs_with_special_tokens(SCREAMING_SNAKE_CASE__ )
a_ : Union[str, Any] = tokenizer.build_inputs_with_special_tokens(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ )
# 2 is for "[CLS]", 3 is for "[SEP]"
assert encoded_sentence == [2] + text + [3]
assert encoded_pair == [2] + text + [3] + text_a + [3]
@custom_tokenizers
class SCREAMING_SNAKE_CASE__ ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE ( self : str ) -> List[str]:
a_ : int = 'cl-tohoku/bert-base-japanese'
a_ : Optional[Any] = AutoTokenizer.from_pretrained(SCREAMING_SNAKE_CASE__ )
self.assertIsInstance(SCREAMING_SNAKE_CASE__ , SCREAMING_SNAKE_CASE__ )
class SCREAMING_SNAKE_CASE__ ( unittest.TestCase ):
def SCREAMING_SNAKE_CASE ( self : str ) -> Optional[Any]:
a_ : List[str] = 'cl-tohoku/bert-base-japanese'
with self.assertLogs('transformers' , level='WARNING' ) as cm:
BertTokenizer.from_pretrained(SCREAMING_SNAKE_CASE__ )
self.assertTrue(
cm.records[0].message.startswith(
'The tokenizer class you load from this checkpoint is not the same type as the class this function'
' is called from.' ) )
a_ : Tuple = 'bert-base-cased'
with self.assertLogs('transformers' , level='WARNING' ) as cm:
BertJapaneseTokenizer.from_pretrained(SCREAMING_SNAKE_CASE__ )
self.assertTrue(
cm.records[0].message.startswith(
'The tokenizer class you load from this checkpoint is not the same type as the class this function'
' is called from.' ) )
| 32 |
"""simple docstring"""
import gc
import math
import unittest
import torch
from diffusers import UNetaDModel
from diffusers.utils import floats_tensor, logging, slow, torch_all_close, torch_device
from diffusers.utils.testing_utils import enable_full_determinism
from .test_modeling_common import ModelTesterMixin, UNetTesterMixin
__a = logging.get_logger(__name__)
enable_full_determinism()
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[Any] = UNetaDModel
_A : Union[str, Any] = """sample"""
@property
def lowerCAmelCase_ ( self: str ) -> Tuple:
snake_case_ :List[str] = 4
snake_case_ :Tuple = 3
snake_case_ :Optional[Any] = (32, 32)
snake_case_ :str = floats_tensor((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :Union[str, Any] = torch.tensor([10] ).to(snake_case )
return {"sample": noise, "timestep": time_step}
@property
def lowerCAmelCase_ ( self: List[str] ) -> Dict:
return (3, 32, 32)
@property
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
return (3, 32, 32)
def lowerCAmelCase_ ( self: Optional[int] ) -> Dict:
snake_case_ :Any = {
"""block_out_channels""": (32, 64),
"""down_block_types""": ("""DownBlock2D""", """AttnDownBlock2D"""),
"""up_block_types""": ("""AttnUpBlock2D""", """UpBlock2D"""),
"""attention_head_dim""": 3,
"""out_channels""": 3,
"""in_channels""": 3,
"""layers_per_block""": 2,
"""sample_size""": 32,
}
snake_case_ :Tuple = self.dummy_input
return init_dict, inputs_dict
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[str] = UNetaDModel
_A : Union[str, Any] = """sample"""
@property
def lowerCAmelCase_ ( self: str ) -> str:
snake_case_ :List[str] = 4
snake_case_ :Optional[int] = 4
snake_case_ :int = (32, 32)
snake_case_ :Any = floats_tensor((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :List[Any] = torch.tensor([10] ).to(snake_case )
return {"sample": noise, "timestep": time_step}
@property
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
return (4, 32, 32)
@property
def lowerCAmelCase_ ( self: List[Any] ) -> int:
return (4, 32, 32)
def lowerCAmelCase_ ( self: Union[str, Any] ) -> List[Any]:
snake_case_ :Dict = {
"""sample_size""": 32,
"""in_channels""": 4,
"""out_channels""": 4,
"""layers_per_block""": 2,
"""block_out_channels""": (32, 64),
"""attention_head_dim""": 32,
"""down_block_types""": ("""DownBlock2D""", """DownBlock2D"""),
"""up_block_types""": ("""UpBlock2D""", """UpBlock2D"""),
}
snake_case_ :List[str] = self.dummy_input
return init_dict, inputs_dict
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[Any]:
snake_case_, snake_case_ :List[str] = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case )
self.assertIsNotNone(snake_case )
self.assertEqual(len(loading_info["""missing_keys"""] ) , 0 )
model.to(snake_case )
snake_case_ :List[str] = model(**self.dummy_input ).sample
assert image is not None, "Make sure output is not None"
@unittest.skipIf(torch_device != """cuda""" , """This test is supposed to run on GPU""" )
def lowerCAmelCase_ ( self: Tuple ) -> Dict:
snake_case_, snake_case_ :Union[str, Any] = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case )
model.to(snake_case )
snake_case_ :Union[str, Any] = model(**self.dummy_input ).sample
assert image is not None, "Make sure output is not None"
@unittest.skipIf(torch_device != """cuda""" , """This test is supposed to run on GPU""" )
def lowerCAmelCase_ ( self: str ) -> Any:
# by defautl model loading will use accelerate as `low_cpu_mem_usage=True`
snake_case_, snake_case_ :List[str] = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case )
model_accelerate.to(snake_case )
model_accelerate.eval()
snake_case_ :List[Any] = torch.randn(
1 , model_accelerate.config.in_channels , model_accelerate.config.sample_size , model_accelerate.config.sample_size , generator=torch.manual_seed(0 ) , )
snake_case_ :int = noise.to(snake_case )
snake_case_ :str = torch.tensor([10] * noise.shape[0] ).to(snake_case )
snake_case_ :Optional[int] = model_accelerate(snake_case , snake_case )["""sample"""]
# two models don't need to stay in the device at the same time
del model_accelerate
torch.cuda.empty_cache()
gc.collect()
snake_case_, snake_case_ :str = UNetaDModel.from_pretrained(
"""fusing/unet-ldm-dummy-update""" , output_loading_info=snake_case , low_cpu_mem_usage=snake_case )
model_normal_load.to(snake_case )
model_normal_load.eval()
snake_case_ :int = model_normal_load(snake_case , snake_case )["""sample"""]
assert torch_all_close(snake_case , snake_case , rtol=1E-3 )
def lowerCAmelCase_ ( self: Tuple ) -> Any:
snake_case_ :Tuple = UNetaDModel.from_pretrained("""fusing/unet-ldm-dummy-update""" )
model.eval()
model.to(snake_case )
snake_case_ :Optional[int] = torch.randn(
1 , model.config.in_channels , model.config.sample_size , model.config.sample_size , generator=torch.manual_seed(0 ) , )
snake_case_ :int = noise.to(snake_case )
snake_case_ :List[Any] = torch.tensor([10] * noise.shape[0] ).to(snake_case )
with torch.no_grad():
snake_case_ :Union[str, Any] = model(snake_case , snake_case ).sample
snake_case_ :Optional[int] = output[0, -1, -3:, -3:].flatten().cpu()
# fmt: off
snake_case_ :Dict = torch.tensor([-1_3.3_2_5_8, -2_0.1_1_0_0, -1_5.9_8_7_3, -1_7.6_6_1_7, -2_3.0_5_9_6, -1_7.9_4_1_9, -1_3.3_6_7_5, -1_6.1_8_8_9, -1_2.3_8_0_0] )
# fmt: on
self.assertTrue(torch_all_close(snake_case , snake_case , rtol=1E-3 ) )
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : List[Any] = UNetaDModel
_A : List[Any] = """sample"""
@property
def lowerCAmelCase_ ( self: Union[str, Any] , snake_case: int=(32, 32) ) -> Tuple:
snake_case_ :Union[str, Any] = 4
snake_case_ :Any = 3
snake_case_ :int = floats_tensor((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :Any = torch.tensor(batch_size * [10] ).to(dtype=torch.intaa , device=snake_case )
return {"sample": noise, "timestep": time_step}
@property
def lowerCAmelCase_ ( self: Union[str, Any] ) -> Any:
return (3, 32, 32)
@property
def lowerCAmelCase_ ( self: int ) -> Tuple:
return (3, 32, 32)
def lowerCAmelCase_ ( self: List[str] ) -> Tuple:
snake_case_ :List[Any] = {
"""block_out_channels""": [32, 64, 64, 64],
"""in_channels""": 3,
"""layers_per_block""": 1,
"""out_channels""": 3,
"""time_embedding_type""": """fourier""",
"""norm_eps""": 1E-6,
"""mid_block_scale_factor""": math.sqrt(2.0 ),
"""norm_num_groups""": None,
"""down_block_types""": [
"""SkipDownBlock2D""",
"""AttnSkipDownBlock2D""",
"""SkipDownBlock2D""",
"""SkipDownBlock2D""",
],
"""up_block_types""": [
"""SkipUpBlock2D""",
"""SkipUpBlock2D""",
"""AttnSkipUpBlock2D""",
"""SkipUpBlock2D""",
],
}
snake_case_ :int = self.dummy_input
return init_dict, inputs_dict
@slow
def lowerCAmelCase_ ( self: Optional[Any] ) -> List[Any]:
snake_case_, snake_case_ :List[Any] = UNetaDModel.from_pretrained("""google/ncsnpp-celebahq-256""" , output_loading_info=snake_case )
self.assertIsNotNone(snake_case )
self.assertEqual(len(loading_info["""missing_keys"""] ) , 0 )
model.to(snake_case )
snake_case_ :Any = self.dummy_input
snake_case_ :int = floats_tensor((4, 3) + (256, 256) ).to(snake_case )
snake_case_ :int = noise
snake_case_ :int = model(**snake_case )
assert image is not None, "Make sure output is not None"
@slow
def lowerCAmelCase_ ( self: str ) -> Dict:
snake_case_ :Dict = UNetaDModel.from_pretrained("""google/ncsnpp-celebahq-256""" )
model.to(snake_case )
snake_case_ :List[str] = 4
snake_case_ :Optional[int] = 3
snake_case_ :List[str] = (256, 256)
snake_case_ :Tuple = torch.ones((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :str = torch.tensor(batch_size * [1E-4] ).to(snake_case )
with torch.no_grad():
snake_case_ :Dict = model(snake_case , snake_case ).sample
snake_case_ :int = output[0, -3:, -3:, -1].flatten().cpu()
# fmt: off
snake_case_ :Optional[Any] = torch.tensor([-4_8_4_2.8_6_9_1, -6_4_9_9.6_6_3_1, -3_8_0_0.1_9_5_3, -7_9_7_8.2_6_8_6, -1_0_9_8_0.7_1_2_9, -2_0_0_2_8.8_5_3_5, 8_1_4_8.2_8_2_2, 2_3_4_2.2_9_0_5, 5_6_7.7_6_0_8] )
# fmt: on
self.assertTrue(torch_all_close(snake_case , snake_case , rtol=1E-2 ) )
def lowerCAmelCase_ ( self: List[str] ) -> List[Any]:
snake_case_ :Optional[Any] = UNetaDModel.from_pretrained("""fusing/ncsnpp-ffhq-ve-dummy-update""" )
model.to(snake_case )
snake_case_ :Optional[int] = 4
snake_case_ :Optional[Any] = 3
snake_case_ :Optional[Any] = (32, 32)
snake_case_ :Dict = torch.ones((batch_size, num_channels) + sizes ).to(snake_case )
snake_case_ :Any = torch.tensor(batch_size * [1E-4] ).to(snake_case )
with torch.no_grad():
snake_case_ :str = model(snake_case , snake_case ).sample
snake_case_ :int = output[0, -3:, -3:, -1].flatten().cpu()
# fmt: off
snake_case_ :int = torch.tensor([-0.0_3_2_5, -0.0_9_0_0, -0.0_8_6_9, -0.0_3_3_2, -0.0_7_2_5, -0.0_2_7_0, -0.0_1_0_1, 0.0_2_2_7, 0.0_2_5_6] )
# fmt: on
self.assertTrue(torch_all_close(snake_case , snake_case , rtol=1E-2 ) )
def lowerCAmelCase_ ( self: Dict ) -> Optional[Any]:
# not required for this model
pass
| 66 | 0 |
"""simple docstring"""
from __future__ import annotations
import requests
__A : Optional[Any] = set(
'''approved_at_utc approved_by author_flair_background_color
author_flair_css_class author_flair_richtext author_flair_template_id author_fullname
author_premium can_mod_post category clicked content_categories created_utc downs
edited gilded gildings hidden hide_score is_created_from_ads_ui is_meta
is_original_content is_reddit_media_domain is_video link_flair_css_class
link_flair_richtext link_flair_text link_flair_text_color media_embed mod_reason_title
name permalink pwls quarantine saved score secure_media secure_media_embed selftext
subreddit subreddit_name_prefixed subreddit_type thumbnail title top_awarded_type
total_awards_received ups upvote_ratio url user_reports'''.split()
)
def lowercase ( __snake_case : str , __snake_case : int = 1 , __snake_case : str = "new" , __snake_case : list | None = None ):
lowercase_ : Tuple = wanted_data or []
if invalid_search_terms := ", ".join(sorted(set(__snake_case ) - valid_terms ) ):
lowercase_ : Union[str, Any] = F'''Invalid search term: {invalid_search_terms}'''
raise ValueError(__snake_case )
lowercase_ : Optional[Any] = requests.get(
F'''https://reddit.com/r/{subreddit}/{age}.json?limit={limit}''' , headers={'''User-agent''': '''A random string'''} , )
if response.status_code == 4_2_9:
raise requests.HTTPError
lowercase_ : Optional[int] = response.json()
if not wanted_data:
return {id_: data["data"]["children"][id_] for id_ in range(__snake_case )}
lowercase_ : str = {}
for id_ in range(__snake_case ):
lowercase_ : Dict = {
item: data['''data''']['''children'''][id_]['''data'''][item] for item in wanted_data
}
return data_dict
if __name__ == "__main__":
# If you get Error 429, that means you are rate limited.Try after some time
print(get_subreddit_data('''learnpython''', wanted_data=['''title''', '''url''', '''selftext''']))
| 33 |
"""simple docstring"""
from typing import TYPE_CHECKING
from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available, is_vision_available
__a = {
"configuration_mask2former": [
"MASK2FORMER_PRETRAINED_CONFIG_ARCHIVE_MAP",
"Mask2FormerConfig",
],
}
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = ["Mask2FormerImageProcessor"]
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
__a = [
"MASK2FORMER_PRETRAINED_MODEL_ARCHIVE_LIST",
"Mask2FormerForUniversalSegmentation",
"Mask2FormerModel",
"Mask2FormerPreTrainedModel",
]
if TYPE_CHECKING:
from .configuration_maskaformer import MASK2FORMER_PRETRAINED_CONFIG_ARCHIVE_MAP, MaskaFormerConfig
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .image_processing_maskaformer import MaskaFormerImageProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_maskaformer import (
MASK2FORMER_PRETRAINED_MODEL_ARCHIVE_LIST,
MaskaFormerForUniversalSegmentation,
MaskaFormerModel,
MaskaFormerPreTrainedModel,
)
else:
import sys
__a = _LazyModule(__name__, globals()["__file__"], _import_structure)
| 66 | 0 |
'''simple docstring'''
from typing import Optional, Union
import numpy as np
from ...image_processing_utils import BaseImageProcessor, BatchFeature
from ...image_transforms import get_image_size, pad, rescale, to_channel_dimension_format
from ...image_utils import ChannelDimension, ImageInput, make_list_of_images, to_numpy_array, valid_images
from ...utils import TensorType, logging
A =logging.get_logger(__name__)
class _a ( __a ):
__a : Dict = ["""pixel_values"""]
def __init__( self : int , lowercase : bool = True , lowercase : Union[int, float] = 1 / 255 , lowercase : bool = True , lowercase : int = 8 , **lowercase : Optional[int] , ):
'''simple docstring'''
super().__init__(**lowercase )
UpperCAmelCase = do_rescale
UpperCAmelCase = rescale_factor
UpperCAmelCase = do_pad
UpperCAmelCase = pad_size
def A ( self : Any , lowercase : np.ndarray , lowercase : float , lowercase : Optional[Union[str, ChannelDimension]] = None , **lowercase : str ):
'''simple docstring'''
return rescale(lowercase , scale=lowercase , data_format=lowercase , **lowercase )
def A ( self : Optional[int] , lowercase : np.ndarray , lowercase : int , lowercase : Optional[Union[str, ChannelDimension]] = None ):
'''simple docstring'''
UpperCAmelCase , UpperCAmelCase = get_image_size(lowercase )
UpperCAmelCase = (old_height // size + 1) * size - old_height
UpperCAmelCase = (old_width // size + 1) * size - old_width
return pad(lowercase , ((0, pad_height), (0, pad_width)) , mode='''symmetric''' , data_format=lowercase )
def A ( self : Union[str, Any] , lowercase : ImageInput , lowercase : Optional[bool] = None , lowercase : Optional[float] = None , lowercase : Optional[bool] = None , lowercase : Optional[int] = None , lowercase : Optional[Union[str, TensorType]] = None , lowercase : Union[str, ChannelDimension] = ChannelDimension.FIRST , **lowercase : Optional[Any] , ):
'''simple docstring'''
UpperCAmelCase = do_rescale if do_rescale is not None else self.do_rescale
UpperCAmelCase = rescale_factor if rescale_factor is not None else self.rescale_factor
UpperCAmelCase = do_pad if do_pad is not None else self.do_pad
UpperCAmelCase = pad_size if pad_size is not None else self.pad_size
UpperCAmelCase = make_list_of_images(lowercase )
if not valid_images(lowercase ):
raise ValueError(
'''Invalid image type. Must be of type PIL.Image.Image, numpy.ndarray, '''
'''torch.Tensor, tf.Tensor or jax.ndarray.''' )
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.
UpperCAmelCase = [to_numpy_array(lowercase ) for image in images]
if do_rescale:
UpperCAmelCase = [self.rescale(image=lowercase , scale=lowercase ) for image in images]
if do_pad:
UpperCAmelCase = [self.pad(lowercase , size=lowercase ) for image in images]
UpperCAmelCase = [to_channel_dimension_format(lowercase , lowercase ) for image in images]
UpperCAmelCase = {'''pixel_values''': images}
return BatchFeature(data=lowercase , tensor_type=lowercase )
| 34 |
"""simple docstring"""
import gc
import unittest
import numpy as np
import torch
from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer
from diffusers import (
AutoencoderKL,
DDIMScheduler,
StableDiffusionSAGPipeline,
UNetaDConditionModel,
)
from diffusers.utils import slow, torch_device
from diffusers.utils.testing_utils import enable_full_determinism, require_torch_gpu
from ..pipeline_params import TEXT_TO_IMAGE_BATCH_PARAMS, TEXT_TO_IMAGE_IMAGE_PARAMS, TEXT_TO_IMAGE_PARAMS
from ..test_pipelines_common import PipelineLatentTesterMixin, PipelineTesterMixin
enable_full_determinism()
class lowerCamelCase ( _lowerCAmelCase , _lowerCAmelCase , unittest.TestCase ):
'''simple docstring'''
_A : str = StableDiffusionSAGPipeline
_A : Optional[Any] = TEXT_TO_IMAGE_PARAMS
_A : Any = TEXT_TO_IMAGE_BATCH_PARAMS
_A : Tuple = TEXT_TO_IMAGE_IMAGE_PARAMS
_A : Tuple = TEXT_TO_IMAGE_IMAGE_PARAMS
_A : List[str] = False
def lowerCAmelCase_ ( self: Optional[Any] ) -> str:
torch.manual_seed(0 )
snake_case_ :Any = UNetaDConditionModel(
block_out_channels=(32, 64) , layers_per_block=2 , sample_size=32 , in_channels=4 , out_channels=4 , down_block_types=("""DownBlock2D""", """CrossAttnDownBlock2D""") , up_block_types=("""CrossAttnUpBlock2D""", """UpBlock2D""") , cross_attention_dim=32 , )
snake_case_ :Any = DDIMScheduler(
beta_start=0.0_0_0_8_5 , beta_end=0.0_1_2 , beta_schedule="""scaled_linear""" , clip_sample=snake_case , set_alpha_to_one=snake_case , )
torch.manual_seed(0 )
snake_case_ :Optional[int] = AutoencoderKL(
block_out_channels=[32, 64] , in_channels=3 , out_channels=3 , down_block_types=["""DownEncoderBlock2D""", """DownEncoderBlock2D"""] , up_block_types=["""UpDecoderBlock2D""", """UpDecoderBlock2D"""] , latent_channels=4 , )
torch.manual_seed(0 )
snake_case_ :Union[str, Any] = CLIPTextConfig(
bos_token_id=0 , eos_token_id=2 , hidden_size=32 , intermediate_size=37 , layer_norm_eps=1E-05 , num_attention_heads=4 , num_hidden_layers=5 , pad_token_id=1 , vocab_size=1_000 , )
snake_case_ :Tuple = CLIPTextModel(snake_case )
snake_case_ :str = CLIPTokenizer.from_pretrained("""hf-internal-testing/tiny-random-clip""" )
snake_case_ :Dict = {
"""unet""": unet,
"""scheduler""": scheduler,
"""vae""": vae,
"""text_encoder""": text_encoder,
"""tokenizer""": tokenizer,
"""safety_checker""": None,
"""feature_extractor""": None,
}
return components
def lowerCAmelCase_ ( self: List[str] , snake_case: Tuple , snake_case: List[str]=0 ) -> str:
if str(snake_case ).startswith("""mps""" ):
snake_case_ :Tuple = torch.manual_seed(snake_case )
else:
snake_case_ :Optional[int] = torch.Generator(device=snake_case ).manual_seed(snake_case )
snake_case_ :Any = {
"""prompt""": """.""",
"""generator""": generator,
"""num_inference_steps""": 2,
"""guidance_scale""": 1.0,
"""sag_scale""": 1.0,
"""output_type""": """numpy""",
}
return inputs
def lowerCAmelCase_ ( self: Optional[int] ) -> str:
super().test_inference_batch_single_identical(expected_max_diff=3E-3 )
@slow
@require_torch_gpu
class lowerCamelCase ( unittest.TestCase ):
'''simple docstring'''
def lowerCAmelCase_ ( self: int ) -> str:
# clean up the VRAM after each test
super().tearDown()
gc.collect()
torch.cuda.empty_cache()
def lowerCAmelCase_ ( self: int ) -> List[str]:
snake_case_ :Any = StableDiffusionSAGPipeline.from_pretrained("""CompVis/stable-diffusion-v1-4""" )
snake_case_ :int = sag_pipe.to(snake_case )
sag_pipe.set_progress_bar_config(disable=snake_case )
snake_case_ :Union[str, Any] = """."""
snake_case_ :str = torch.manual_seed(0 )
snake_case_ :str = sag_pipe(
[prompt] , generator=snake_case , guidance_scale=7.5 , sag_scale=1.0 , num_inference_steps=20 , output_type="""np""" )
snake_case_ :List[Any] = output.images
snake_case_ :Tuple = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
snake_case_ :List[Any] = np.array([0.1_5_6_8, 0.1_7_3_8, 0.1_6_9_5, 0.1_6_9_3, 0.1_5_0_7, 0.1_7_0_5, 0.1_5_4_7, 0.1_7_5_1, 0.1_9_4_9] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 5E-2
def lowerCAmelCase_ ( self: Dict ) -> str:
snake_case_ :Tuple = StableDiffusionSAGPipeline.from_pretrained("""stabilityai/stable-diffusion-2-1-base""" )
snake_case_ :Optional[int] = sag_pipe.to(snake_case )
sag_pipe.set_progress_bar_config(disable=snake_case )
snake_case_ :Tuple = """."""
snake_case_ :Union[str, Any] = torch.manual_seed(0 )
snake_case_ :Tuple = sag_pipe(
[prompt] , generator=snake_case , guidance_scale=7.5 , sag_scale=1.0 , num_inference_steps=20 , output_type="""np""" )
snake_case_ :Optional[int] = output.images
snake_case_ :Optional[Any] = image[0, -3:, -3:, -1]
assert image.shape == (1, 512, 512, 3)
snake_case_ :Tuple = np.array([0.3_4_5_9, 0.2_8_7_6, 0.2_5_3_7, 0.3_0_0_2, 0.2_6_7_1, 0.2_1_6_0, 0.3_0_2_6, 0.2_2_6_2, 0.2_3_7_1] )
assert np.abs(image_slice.flatten() - expected_slice ).max() < 5E-2
def lowerCAmelCase_ ( self: List[str] ) -> List[str]:
snake_case_ :Optional[int] = StableDiffusionSAGPipeline.from_pretrained("""stabilityai/stable-diffusion-2-1-base""" )
snake_case_ :int = sag_pipe.to(snake_case )
sag_pipe.set_progress_bar_config(disable=snake_case )
snake_case_ :Tuple = """."""
snake_case_ :Optional[int] = torch.manual_seed(0 )
snake_case_ :List[str] = sag_pipe(
[prompt] , width=768 , height=512 , generator=snake_case , guidance_scale=7.5 , sag_scale=1.0 , num_inference_steps=20 , output_type="""np""" , )
snake_case_ :Optional[Any] = output.images
assert image.shape == (1, 512, 768, 3)
| 66 | 0 |
'''simple docstring'''
# HF Trainer benchmarking tool
#
# This tool can be used to run and compare multiple dimensions of the HF Trainers args.
#
# It then prints a report once in github format with all the information that needs to be shared
# with others and second time in a console-friendly format, so it's easier to use for tuning things up.
#
# The main idea is:
#
# ./trainer-benchmark.py --base-cmd '<cmd args that don't change>' \
# --variations '--tf32 0|--tf32 1' '--fp16 0|--fp16 1|--bf16 1' \
# --target-metric-key train_samples_per_second
#
# The variations can be any command line argument that you want to compare and not just dtype as in
# the example.
#
# --variations allows you to compare variations in multiple dimensions.
#
# as the first dimention has 2 options and the second 3 in our example, this will run the trainer 6
# times adding one of:
#
# 1. --tf32 0 --fp16 0
# 2. --tf32 0 --fp16 1
# 3. --tf32 0 --bf16 1
# 4. --tf32 1 --fp16 0
# 5. --tf32 1 --fp16 1
# 6. --tf32 1 --bf16 1
#
# and print the results. This is just a cartesian product - and more than 2 dimensions can be used.
#
# If you want to rely on defaults, this:
# --variations '--tf32 0|--tf32 1' '--fp16 0|--fp16 1|--bf16 1'
# is identical to this:
# --variations '--tf32 0|--tf32 1' '|--fp16|--bf16'
#
# the leading empty variation in the 2nd dimension is a valid variation.
#
# So here we get the following 6 variations:
#
# 1. --tf32 0
# 2. --tf32 0 --fp16
# 3. --tf32 0 --bf16
# 4. --tf32 1
# 5. --tf32 1 --fp16
# 6. --tf32 1 --bf16
#
# In this particular case we don't know what the default tf32 setting is as it's normally
# pytorch-version dependent). That's why it's best to do an explicit setting of each variation:
# `--tf32 0|--tf32 1`
#
# Here is a full example of a train:
#
# CUDA_VISIBLE_DEVICES=0 python ./scripts/benchmark/trainer-benchmark.py \
# --base-cmd \
# ' examples/pytorch/translation/run_translation.py --model_name_or_path t5-small \
# --output_dir output_dir --do_train --label_smoothing 0.1 --logging_strategy no \
# --save_strategy no --per_device_train_batch_size 32 --max_source_length 512 \
# --max_target_length 512 --num_train_epochs 1 --overwrite_output_dir \
# --source_lang en --target_lang ro --dataset_name wmt16 --dataset_config "ro-en" \
# --source_prefix "translate English to Romanian: " --warmup_steps 50 \
# --max_train_samples 20000 --dataloader_num_workers 2 ' \
# --target-metric-key train_samples_per_second --repeat-times 1 --variations \
# '|--fp16|--bf16' '--tf32 0|--tf32 1' --report-metric-keys train_loss \
# --repeat-times 1 --base-variation '--tf32 0'
#
# and here is a possible output:
#
#
# | Variation | Train | Diff | Train |
# | | samples | % | loss |
# | | per | | |
# | | second | | |
# |:----------------|----------:|-------:|--------:|
# | --tf32 0 | 285.11 | 0 | 2.51 |
# | --tf32 1 | 342.09 | 20 | 2.51 |
# | --fp16 --tf32 0 | 423.49 | 49 | 2.51 |
# | --fp16 --tf32 1 | 423.13 | 48 | 2.51 |
# | --bf16 --tf32 0 | 416.80 | 46 | 2.52 |
# | --bf16 --tf32 1 | 415.87 | 46 | 2.52 |
#
#
# So you can quickly compare the different outcomes.
#
# Typically running each experiment once is enough, but if the environment is unstable you can
# re-run each multiple times, e.g., 3 using --repeat-times 3 and it will report the averaged results.
#
# By default it'll use the lowest result as the base line to use as 100% and then compare the rest to
# it as can be seen from the table above, but you can also specify which combination is the one to use as
# the baseline, e.g., to change to another entry use: --base-variation '--tf32 1 --fp16 0'
#
# --target-metric-key is there to tell the program which metrics to compare - the different metric keys are
# inside output_dir/all_results.json. e.g., to measure eval performance instead of train use:
# --target-metric-key eval_samples_per_second
# but of course you will need to adjust the --base-cmd value in the example to perform evaluation as
# well (as currently it doesn't)
#
import argparse
import datetime
import io
import itertools
import json
import math
import os
import platform
import re
import shlex
import subprocess
import sys
from pathlib import Path
from statistics import fmean
import pandas as pd
import torch
from tqdm import tqdm
import transformers
__a = float("nan")
class UpperCAmelCase_ :
"""simple docstring"""
def __init__( self : str , snake_case_ : str ):
snake_case__ : Union[str, Any] = sys.stdout
snake_case__ : int = open(snake_case_ , """a""" )
def __getattr__( self : Tuple , snake_case_ : Optional[Any] ):
return getattr(self.stdout , snake_case_ )
def lowerCamelCase ( self : Any , snake_case_ : List[str] ):
self.stdout.write(snake_case_ )
# strip tqdm codes
self.file.write(re.sub(r"""^.*\r""" , """""" , snake_case_ , 0 , re.M ) )
def __snake_case( _lowerCAmelCase=80 , _lowerCAmelCase=False ) -> List[str]:
snake_case__ : Union[str, Any] = []
# deal with critical env vars
snake_case__ : Optional[int] = ["""CUDA_VISIBLE_DEVICES"""]
for key in env_keys:
snake_case__ : List[str] = os.environ.get(_lowerCAmelCase , _lowerCAmelCase )
if val is not None:
cmd.append(f"{key}={val}" )
# python executable (not always needed if the script is executable)
snake_case__ : Optional[Any] = sys.executable if full_python_path else sys.executable.split("""/""" )[-1]
cmd.append(_lowerCAmelCase )
# now the normal args
cmd += list(map(shlex.quote , sys.argv ) )
# split up into up to MAX_WIDTH lines with shell multi-line escapes
snake_case__ : List[Any] = []
snake_case__ : Any = """"""
while len(_lowerCAmelCase ) > 0:
current_line += f"{cmd.pop(0 )} "
if len(_lowerCAmelCase ) == 0 or len(_lowerCAmelCase ) + len(cmd[0] ) + 1 > max_width - 1:
lines.append(_lowerCAmelCase )
snake_case__ : int = """"""
return "\\\n".join(_lowerCAmelCase )
def __snake_case( _lowerCAmelCase , _lowerCAmelCase ) -> int:
# unwrap multi-line input
snake_case__ : Dict = re.sub(r"""[\\\n]+""" , """ """ , args.base_cmd )
# remove --output_dir if any and set our own
snake_case__ : Any = re.sub("""--output_dir\s+[^\s]+""" , """""" , args.base_cmd )
args.base_cmd += f" --output_dir {output_dir}"
# ensure we have --overwrite_output_dir
snake_case__ : str = re.sub("""--overwrite_output_dir\s+""" , """""" , args.base_cmd )
args.base_cmd += " --overwrite_output_dir"
return [sys.executable] + shlex.split(args.base_cmd )
def __snake_case( _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase ) -> Optional[Any]:
# Enable to debug everything but the run itself, to do it fast and see the progress.
# This is useful for debugging the output formatting quickly - we can remove it later once
# everybody is happy with the output
if 0:
import random
from time import sleep
sleep(0 )
return dict(
{k: random.uniform(0 , 100 ) for k in metric_keys} , **{target_metric_key: random.choice([nan, 10.31, 100.2, 55.6666, 222.2222_2222] )} , )
snake_case__ : Union[str, Any] = subprocess.run(_lowerCAmelCase , capture_output=_lowerCAmelCase , text=_lowerCAmelCase )
if verbose:
print("""STDOUT""" , result.stdout )
print("""STDERR""" , result.stderr )
# save the streams
snake_case__ : Dict = variation.replace(""" """ , """-""" )
with open(Path(_lowerCAmelCase ) / f"log.{prefix}.stdout.txt" , """w""" ) as f:
f.write(result.stdout )
with open(Path(_lowerCAmelCase ) / f"log.{prefix}.stderr.txt" , """w""" ) as f:
f.write(result.stderr )
if result.returncode != 0:
if verbose:
print("""failed""" )
return {target_metric_key: nan}
with io.open(f"{output_dir}/all_results.json" , """r""" , encoding="""utf-8""" ) as f:
snake_case__ : Dict = json.load(_lowerCAmelCase )
# filter out just the keys we want
return {k: v for k, v in metrics.items() if k in metric_keys}
def __snake_case( _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , ) -> Dict:
snake_case__ : Any = []
snake_case__ : int = []
snake_case__ : Tuple = f"{id}: {variation:<{longest_variation_len}}"
snake_case__ : Optional[Any] = f"{preamble}: "
snake_case__ : Optional[Any] = set(report_metric_keys + [target_metric_key] )
for i in tqdm(range(_lowerCAmelCase ) , desc=_lowerCAmelCase , leave=_lowerCAmelCase ):
snake_case__ : Dict = process_run_single(
_lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase )
snake_case__ : Tuple = single_run_metrics[target_metric_key]
if not math.isnan(_lowerCAmelCase ):
metrics.append(_lowerCAmelCase )
results.append(_lowerCAmelCase )
outcome += "✓"
else:
outcome += "✘"
snake_case__ : str = f"\33[2K\r{outcome}"
if len(_lowerCAmelCase ) > 0:
snake_case__ : List[Any] = {k: fmean([x[k] for x in metrics] ) for k in metrics[0].keys()}
snake_case__ : Any = round(mean_metrics[target_metric_key] , 2 )
snake_case__ : Optional[Any] = f"{outcome} {mean_target}"
if len(_lowerCAmelCase ) > 1:
results_str += f" {tuple(round(_lowerCAmelCase , 2 ) for x in results )}"
print(_lowerCAmelCase )
snake_case__ : Optional[Any] = variation
return mean_metrics
else:
print(_lowerCAmelCase )
return {variation_key: variation, target_metric_key: nan}
def __snake_case( ) -> Any:
snake_case__ : int = torch.cuda.get_device_properties(torch.device("""cuda""" ) )
return f"\nDatetime : {datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S' )}\n\nSoftware:\ntransformers: {transformers.__version__}\ntorch : {torch.__version__}\ncuda : {torch.version.cuda}\npython : {platform.python_version()}\n\nHardware:\n{torch.cuda.device_count()} GPUs : {properties.name}, {properties.total_memory/2**30:0.2f}GB\n"
def __snake_case( _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase ) -> Optional[int]:
snake_case__ : List[Any] = pd.DataFrame(_lowerCAmelCase )
snake_case__ : Union[str, Any] = """variation"""
snake_case__ : int = """diff_%"""
snake_case__ : List[Any] = nan
if base_variation is not None and len(df[df[variation_key] == base_variation] ):
# this may still return nan
snake_case__ : Tuple = df.loc[df[variation_key] == base_variation][target_metric_key].item()
if math.isnan(_lowerCAmelCase ):
# as a fallback, use the minimal value as the sentinel
snake_case__ : Optional[int] = df.loc[df[target_metric_key] != nan][target_metric_key].min()
# create diff column if possible
if not math.isnan(_lowerCAmelCase ):
snake_case__ : Optional[Any] = df.apply(
lambda _lowerCAmelCase : round(100 * (r[target_metric_key] - sentinel_value) / sentinel_value )
if not math.isnan(r[target_metric_key] )
else 0 , axis="""columns""" , )
# re-order columns
snake_case__ : str = [variation_key, target_metric_key, diff_key, *report_metric_keys]
snake_case__ : int = df.reindex(_lowerCAmelCase , axis="""columns""" ) # reorder cols
# capitalize
snake_case__ : Any = df.rename(str.capitalize , axis="""columns""" )
# make the cols as narrow as possible
snake_case__ : Optional[Any] = df.rename(lambda _lowerCAmelCase : c.replace("""_""" , """<br>""" ) , axis="""columns""" )
snake_case__ : Optional[Any] = df.rename(lambda _lowerCAmelCase : c.replace("""_""" , """\n""" ) , axis="""columns""" )
snake_case__ : Optional[int] = ["""""", """Copy between the cut-here-lines and paste as is to github or a forum"""]
report += ["----------8<-----------------8<--------"]
report += ["*** Results:", df_github.to_markdown(index=_lowerCAmelCase , floatfmt=""".2f""" )]
report += ["```"]
report += ["*** Setup:", get_versions()]
report += ["*** The benchmark command line was:", get_original_command()]
report += ["```"]
report += ["----------8<-----------------8<--------"]
report += ["*** Results (console):", df_console.to_markdown(index=_lowerCAmelCase , floatfmt=""".2f""" )]
print("""\n\n""".join(_lowerCAmelCase ) )
def __snake_case( ) -> Any:
snake_case__ : int = argparse.ArgumentParser()
parser.add_argument(
"""--base-cmd""" , default=_lowerCAmelCase , type=_lowerCAmelCase , required=_lowerCAmelCase , help="""Base cmd""" , )
parser.add_argument(
"""--variations""" , default=_lowerCAmelCase , type=_lowerCAmelCase , nargs="""+""" , required=_lowerCAmelCase , help="""Multi-dimensional variations, example: '|--fp16|--bf16' '|--tf32'""" , )
parser.add_argument(
"""--base-variation""" , default=_lowerCAmelCase , type=_lowerCAmelCase , help="""Baseline variation to compare to. if None the minimal target value will be used to compare against""" , )
parser.add_argument(
"""--target-metric-key""" , default=_lowerCAmelCase , type=_lowerCAmelCase , required=_lowerCAmelCase , help="""Target metric key in output_dir/all_results.json, e.g., train_samples_per_second""" , )
parser.add_argument(
"""--report-metric-keys""" , default="""""" , type=_lowerCAmelCase , help="""Report metric keys - other metric keys from output_dir/all_results.json to report, e.g., train_loss. Use a single argument e.g., 'train_loss train_samples""" , )
parser.add_argument(
"""--repeat-times""" , default=1 , type=_lowerCAmelCase , help="""How many times to re-run each variation - an average will be reported""" , )
parser.add_argument(
"""--output_dir""" , default="""output_benchmark""" , type=_lowerCAmelCase , help="""The output directory where all the benchmark reports will go to and additionally this directory will be used to override --output_dir in the script that is being benchmarked""" , )
parser.add_argument(
"""--verbose""" , default=_lowerCAmelCase , action="""store_true""" , help="""Whether to show the outputs of each run or just the benchmark progress""" , )
snake_case__ : int = parser.parse_args()
snake_case__ : Dict = args.output_dir
Path(_lowerCAmelCase ).mkdir(exist_ok=_lowerCAmelCase )
snake_case__ : Dict = get_base_command(_lowerCAmelCase , _lowerCAmelCase )
# split each dimension into its --foo variations
snake_case__ : Dict = [list(map(str.strip , re.split(r"""\|""" , _lowerCAmelCase ) ) ) for x in args.variations]
# build a cartesian product of dimensions and convert those back into cmd-line arg strings,
# while stripping white space for inputs that were empty
snake_case__ : List[str] = list(map(str.strip , map(""" """.join , itertools.product(*_lowerCAmelCase ) ) ) )
snake_case__ : List[str] = max(len(_lowerCAmelCase ) for x in variations )
# split wanted keys
snake_case__ : int = args.report_metric_keys.split()
# capture prints into a log file for convenience
snake_case__ : str = f"benchmark-report-{datetime.datetime.now().strftime('%Y-%m-%d-%H-%M-%S' )}.txt"
print(f"\nNote: each run's output is also logged under {output_dir}/log.*.std*.txt" )
print(f"and this script's output is also piped into {report_fn}" )
snake_case__ : Optional[int] = Tee(_lowerCAmelCase )
print(f"\n*** Running {len(_lowerCAmelCase )} benchmarks:" )
print(f"Base command: {' '.join(_lowerCAmelCase )}" )
snake_case__ : Any = """variation"""
snake_case__ : str = []
for id, variation in enumerate(tqdm(_lowerCAmelCase , desc="""Total completion: """ , leave=_lowerCAmelCase ) ):
snake_case__ : str = base_cmd + variation.split()
results.append(
process_run(
id + 1 , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , _lowerCAmelCase , args.target_metric_key , _lowerCAmelCase , args.repeat_times , _lowerCAmelCase , args.verbose , ) )
process_results(_lowerCAmelCase , args.target_metric_key , _lowerCAmelCase , args.base_variation , _lowerCAmelCase )
if __name__ == "__main__":
main()
| 35 |
"""simple docstring"""
from __future__ import annotations
from collections import Counter
from random import random
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Tuple ) -> Optional[Any]:
snake_case_ :Optional[int] = {}
def lowerCAmelCase_ ( self: Dict , snake_case: str ) -> None:
snake_case_ :str = {}
def lowerCAmelCase_ ( self: Optional[int] , snake_case: str , snake_case: str , snake_case: float ) -> None:
if nodea not in self.connections:
self.add_node(snake_case )
if nodea not in self.connections:
self.add_node(snake_case )
snake_case_ :Dict = probability
def lowerCAmelCase_ ( self: List[Any] ) -> list[str]:
return list(self.connections )
def lowerCAmelCase_ ( self: Any , snake_case: str ) -> str:
snake_case_ :Optional[Any] = 0
snake_case_ :List[str] = random()
for dest in self.connections[node]:
current_probability += self.connections[node][dest]
if current_probability > random_value:
return dest
return ""
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :List[str] = MarkovChainGraphUndirectedUnweighted()
for nodea, nodea, probability in transitions:
graph.add_transition_probability(_lowercase, _lowercase, _lowercase )
snake_case_ :int = Counter(graph.get_nodes() )
snake_case_ :Optional[Any] = start
for _ in range(_lowercase ):
snake_case_ :Tuple = graph.transition(_lowercase )
visited[node] += 1
return visited
if __name__ == "__main__":
import doctest
doctest.testmod()
| 66 | 0 |
from typing import Any, Callable, Dict, List, Optional, Union
import torch
from transformers import CLIPImageProcessor, CLIPTextModel, CLIPTokenizer
from diffusers import (
AutoencoderKL,
DDIMScheduler,
DiffusionPipeline,
LMSDiscreteScheduler,
PNDMScheduler,
StableDiffusionPipeline,
UNetaDConditionModel,
)
from diffusers.pipelines.stable_diffusion import StableDiffusionPipelineOutput
from diffusers.pipelines.stable_diffusion.safety_checker import StableDiffusionSafetyChecker
_snake_case = "CompVis/stable-diffusion-v1-1"
_snake_case = "CompVis/stable-diffusion-v1-2"
_snake_case = "CompVis/stable-diffusion-v1-3"
_snake_case = "CompVis/stable-diffusion-v1-4"
class UpperCAmelCase_ ( a):
def __init__( self, __a, __a, __a, __a, __a, __a, __a, __a = True, ):
'''simple docstring'''
super()._init_()
_lowerCAmelCase : List[Any] = StableDiffusionPipeline.from_pretrained(__a)
_lowerCAmelCase : List[Any] = StableDiffusionPipeline.from_pretrained(__a)
_lowerCAmelCase : str = StableDiffusionPipeline.from_pretrained(__a)
_lowerCAmelCase : str = StableDiffusionPipeline(
vae=__a, text_encoder=__a, tokenizer=__a, unet=__a, scheduler=__a, safety_checker=__a, feature_extractor=__a, requires_safety_checker=__a, )
self.register_modules(pipelinea=self.pipea, pipelinea=self.pipea, pipelinea=self.pipea, pipelinea=self.pipea)
@property
def snake_case__ ( self):
'''simple docstring'''
return {k: getattr(self, __a) for k in self.config.keys() if not k.startswith("_")}
def snake_case__ ( self, __a = "auto"):
'''simple docstring'''
if slice_size == "auto":
# half the attention head size is usually a good trade-off between
# speed and memory
_lowerCAmelCase : Union[str, Any] = self.unet.config.attention_head_dim // 2
self.unet.set_attention_slice(__a)
def snake_case__ ( self):
'''simple docstring'''
self.enable_attention_slicing(__a)
@torch.no_grad()
def snake_case__ ( self, __a, __a = 512, __a = 512, __a = 50, __a = 7.5, __a = None, __a = 1, __a = 0.0, __a = None, __a = None, __a = "pil", __a = True, __a = None, __a = 1, **__a, ):
'''simple docstring'''
return self.pipea(
prompt=__a, height=__a, width=__a, num_inference_steps=__a, guidance_scale=__a, negative_prompt=__a, num_images_per_prompt=__a, eta=__a, generator=__a, latents=__a, output_type=__a, return_dict=__a, callback=__a, callback_steps=__a, **__a, )
@torch.no_grad()
def snake_case__ ( self, __a, __a = 512, __a = 512, __a = 50, __a = 7.5, __a = None, __a = 1, __a = 0.0, __a = None, __a = None, __a = "pil", __a = True, __a = None, __a = 1, **__a, ):
'''simple docstring'''
return self.pipea(
prompt=__a, height=__a, width=__a, num_inference_steps=__a, guidance_scale=__a, negative_prompt=__a, num_images_per_prompt=__a, eta=__a, generator=__a, latents=__a, output_type=__a, return_dict=__a, callback=__a, callback_steps=__a, **__a, )
@torch.no_grad()
def snake_case__ ( self, __a, __a = 512, __a = 512, __a = 50, __a = 7.5, __a = None, __a = 1, __a = 0.0, __a = None, __a = None, __a = "pil", __a = True, __a = None, __a = 1, **__a, ):
'''simple docstring'''
return self.pipea(
prompt=__a, height=__a, width=__a, num_inference_steps=__a, guidance_scale=__a, negative_prompt=__a, num_images_per_prompt=__a, eta=__a, generator=__a, latents=__a, output_type=__a, return_dict=__a, callback=__a, callback_steps=__a, **__a, )
@torch.no_grad()
def snake_case__ ( self, __a, __a = 512, __a = 512, __a = 50, __a = 7.5, __a = None, __a = 1, __a = 0.0, __a = None, __a = None, __a = "pil", __a = True, __a = None, __a = 1, **__a, ):
'''simple docstring'''
return self.pipea(
prompt=__a, height=__a, width=__a, num_inference_steps=__a, guidance_scale=__a, negative_prompt=__a, num_images_per_prompt=__a, eta=__a, generator=__a, latents=__a, output_type=__a, return_dict=__a, callback=__a, callback_steps=__a, **__a, )
@torch.no_grad()
def snake_case__ ( self, __a, __a = 512, __a = 512, __a = 50, __a = 7.5, __a = None, __a = 1, __a = 0.0, __a = None, __a = None, __a = "pil", __a = True, __a = None, __a = 1, **__a, ):
'''simple docstring'''
_lowerCAmelCase : Tuple = "cuda" if torch.cuda.is_available() else "cpu"
self.to(__a)
# Checks if the height and width are divisible by 8 or not
if height % 8 != 0 or width % 8 != 0:
raise ValueError(f"`height` and `width` must be divisible by 8 but are {height} and {width}.")
# Get first result from Stable Diffusion Checkpoint v1.1
_lowerCAmelCase : Any = self.textaimg_sda_a(
prompt=__a, height=__a, width=__a, num_inference_steps=__a, guidance_scale=__a, negative_prompt=__a, num_images_per_prompt=__a, eta=__a, generator=__a, latents=__a, output_type=__a, return_dict=__a, callback=__a, callback_steps=__a, **__a, )
# Get first result from Stable Diffusion Checkpoint v1.2
_lowerCAmelCase : int = self.textaimg_sda_a(
prompt=__a, height=__a, width=__a, num_inference_steps=__a, guidance_scale=__a, negative_prompt=__a, num_images_per_prompt=__a, eta=__a, generator=__a, latents=__a, output_type=__a, return_dict=__a, callback=__a, callback_steps=__a, **__a, )
# Get first result from Stable Diffusion Checkpoint v1.3
_lowerCAmelCase : str = self.textaimg_sda_a(
prompt=__a, height=__a, width=__a, num_inference_steps=__a, guidance_scale=__a, negative_prompt=__a, num_images_per_prompt=__a, eta=__a, generator=__a, latents=__a, output_type=__a, return_dict=__a, callback=__a, callback_steps=__a, **__a, )
# Get first result from Stable Diffusion Checkpoint v1.4
_lowerCAmelCase : Union[str, Any] = self.textaimg_sda_a(
prompt=__a, height=__a, width=__a, num_inference_steps=__a, guidance_scale=__a, negative_prompt=__a, num_images_per_prompt=__a, eta=__a, generator=__a, latents=__a, output_type=__a, return_dict=__a, callback=__a, callback_steps=__a, **__a, )
# Get all result images into a single list and pass it via StableDiffusionPipelineOutput for final result
return StableDiffusionPipelineOutput([resa[0], resa[0], resa[0], resa[0]])
| 36 |
"""simple docstring"""
import argparse
import collections
import os
import re
import tempfile
import pandas as pd
from datasets import Dataset
from huggingface_hub import hf_hub_download, upload_folder
from transformers.utils import direct_transformers_import
# All paths are set with the intent you should run this script from the root of the repo with the command
# python utils/update_metadata.py
__a = "src/transformers"
# This is to make sure the transformers module imported is the one in the repo.
__a = direct_transformers_import(TRANSFORMERS_PATH)
# Regexes that match TF/Flax/PT model names.
__a = re.compile(r"TF(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
__a = re.compile(r"Flax(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
# Will match any TF or Flax model too so need to be in an else branch afterthe two previous regexes.
__a = re.compile(r"(.*)(?:Model|Encoder|Decoder|ForConditionalGeneration)")
# Fill this with tuples (pipeline_tag, model_mapping, auto_model)
__a = [
("pretraining", "MODEL_FOR_PRETRAINING_MAPPING_NAMES", "AutoModelForPreTraining"),
("feature-extraction", "MODEL_MAPPING_NAMES", "AutoModel"),
("audio-classification", "MODEL_FOR_AUDIO_CLASSIFICATION_MAPPING_NAMES", "AutoModelForAudioClassification"),
("text-generation", "MODEL_FOR_CAUSAL_LM_MAPPING_NAMES", "AutoModelForCausalLM"),
("automatic-speech-recognition", "MODEL_FOR_CTC_MAPPING_NAMES", "AutoModelForCTC"),
("image-classification", "MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES", "AutoModelForImageClassification"),
("image-segmentation", "MODEL_FOR_IMAGE_SEGMENTATION_MAPPING_NAMES", "AutoModelForImageSegmentation"),
("fill-mask", "MODEL_FOR_MASKED_LM_MAPPING_NAMES", "AutoModelForMaskedLM"),
("object-detection", "MODEL_FOR_OBJECT_DETECTION_MAPPING_NAMES", "AutoModelForObjectDetection"),
(
"zero-shot-object-detection",
"MODEL_FOR_ZERO_SHOT_OBJECT_DETECTION_MAPPING_NAMES",
"AutoModelForZeroShotObjectDetection",
),
("question-answering", "MODEL_FOR_QUESTION_ANSWERING_MAPPING_NAMES", "AutoModelForQuestionAnswering"),
("text2text-generation", "MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING_NAMES", "AutoModelForSeq2SeqLM"),
("text-classification", "MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES", "AutoModelForSequenceClassification"),
("automatic-speech-recognition", "MODEL_FOR_SPEECH_SEQ_2_SEQ_MAPPING_NAMES", "AutoModelForSpeechSeq2Seq"),
(
"table-question-answering",
"MODEL_FOR_TABLE_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForTableQuestionAnswering",
),
("token-classification", "MODEL_FOR_TOKEN_CLASSIFICATION_MAPPING_NAMES", "AutoModelForTokenClassification"),
("multiple-choice", "MODEL_FOR_MULTIPLE_CHOICE_MAPPING_NAMES", "AutoModelForMultipleChoice"),
(
"next-sentence-prediction",
"MODEL_FOR_NEXT_SENTENCE_PREDICTION_MAPPING_NAMES",
"AutoModelForNextSentencePrediction",
),
(
"audio-frame-classification",
"MODEL_FOR_AUDIO_FRAME_CLASSIFICATION_MAPPING_NAMES",
"AutoModelForAudioFrameClassification",
),
("audio-xvector", "MODEL_FOR_AUDIO_XVECTOR_MAPPING_NAMES", "AutoModelForAudioXVector"),
(
"document-question-answering",
"MODEL_FOR_DOCUMENT_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForDocumentQuestionAnswering",
),
(
"visual-question-answering",
"MODEL_FOR_VISUAL_QUESTION_ANSWERING_MAPPING_NAMES",
"AutoModelForVisualQuestionAnswering",
),
("image-to-text", "MODEL_FOR_FOR_VISION_2_SEQ_MAPPING_NAMES", "AutoModelForVision2Seq"),
(
"zero-shot-image-classification",
"MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING_NAMES",
"AutoModelForZeroShotImageClassification",
),
("depth-estimation", "MODEL_FOR_DEPTH_ESTIMATION_MAPPING_NAMES", "AutoModelForDepthEstimation"),
("video-classification", "MODEL_FOR_VIDEO_CLASSIFICATION_MAPPING_NAMES", "AutoModelForVideoClassification"),
("mask-generation", "MODEL_FOR_MASK_GENERATION_MAPPING_NAMES", "AutoModelForMaskGeneration"),
]
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Any = re.finditer(""".+?(?:(?<=[a-z])(?=[A-Z])|(?<=[A-Z])(?=[A-Z][a-z])|$)""", _lowercase )
return [m.group(0 ) for m in matches]
def A_ ( ):
'''simple docstring'''
snake_case_ :int = transformers_module.models.auto.configuration_auto.CONFIG_MAPPING_NAMES
snake_case_ :Dict = {
config.replace("""Config""", """""" ): model_type for model_type, config in config_maping_names.items()
}
# Dictionaries flagging if each model prefix has a backend in PT/TF/Flax.
snake_case_ :Optional[Any] = collections.defaultdict(_lowercase )
snake_case_ :int = collections.defaultdict(_lowercase )
snake_case_ :List[str] = collections.defaultdict(_lowercase )
# Let's lookup through all transformers object (once) and find if models are supported by a given backend.
for attr_name in dir(_lowercase ):
snake_case_ :int = None
if _re_tf_models.match(_lowercase ) is not None:
snake_case_ :int = tf_models
snake_case_ :List[str] = _re_tf_models.match(_lowercase ).groups()[0]
elif _re_flax_models.match(_lowercase ) is not None:
snake_case_ :List[Any] = flax_models
snake_case_ :Any = _re_flax_models.match(_lowercase ).groups()[0]
elif _re_pt_models.match(_lowercase ) is not None:
snake_case_ :Optional[Any] = pt_models
snake_case_ :int = _re_pt_models.match(_lowercase ).groups()[0]
if lookup_dict is not None:
while len(_lowercase ) > 0:
if attr_name in model_prefix_to_model_type:
snake_case_ :Optional[int] = True
break
# Try again after removing the last word in the name
snake_case_ :Optional[Any] = """""".join(camel_case_split(_lowercase )[:-1] )
snake_case_ :Optional[int] = set(list(pt_models.keys() ) + list(tf_models.keys() ) + list(flax_models.keys() ) )
snake_case_ :Optional[Any] = list(_lowercase )
all_models.sort()
snake_case_ :Optional[int] = {"""model_type""": all_models}
snake_case_ :Optional[int] = [pt_models[t] for t in all_models]
snake_case_ :Any = [tf_models[t] for t in all_models]
snake_case_ :Dict = [flax_models[t] for t in all_models]
# Now let's use the auto-mapping names to make sure
snake_case_ :Dict = {}
for t in all_models:
if t in transformers_module.models.auto.processing_auto.PROCESSOR_MAPPING_NAMES:
snake_case_ :Optional[Any] = """AutoProcessor"""
elif t in transformers_module.models.auto.tokenization_auto.TOKENIZER_MAPPING_NAMES:
snake_case_ :Tuple = """AutoTokenizer"""
elif t in transformers_module.models.auto.feature_extraction_auto.FEATURE_EXTRACTOR_MAPPING_NAMES:
snake_case_ :Tuple = """AutoFeatureExtractor"""
else:
# Default to AutoTokenizer if a model has nothing, for backward compatibility.
snake_case_ :str = """AutoTokenizer"""
snake_case_ :int = [processors[t] for t in all_models]
return pd.DataFrame(_lowercase )
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :List[Any] = [
transformers_module.models.auto.modeling_auto,
transformers_module.models.auto.modeling_tf_auto,
transformers_module.models.auto.modeling_flax_auto,
]
for pipeline_tag, model_mapping, auto_class in PIPELINE_TAGS_AND_AUTO_MODELS:
snake_case_ :Optional[int] = [model_mapping, f"""TF_{model_mapping}""", f"""FLAX_{model_mapping}"""]
snake_case_ :List[str] = [auto_class, f"""TF_{auto_class}""", f"""Flax_{auto_class}"""]
# Loop through all three frameworks
for module, cls, mapping in zip(_lowercase, _lowercase, _lowercase ):
# The type of pipeline may not exist in this framework
if not hasattr(_lowercase, _lowercase ):
continue
# First extract all model_names
snake_case_ :Tuple = []
for name in getattr(_lowercase, _lowercase ).values():
if isinstance(_lowercase, _lowercase ):
model_names.append(_lowercase )
else:
model_names.extend(list(_lowercase ) )
# Add pipeline tag and auto model class for those models
table.update({model_name: (pipeline_tag, cls) for model_name in model_names} )
return table
def A_ ( _lowercase, _lowercase ):
'''simple docstring'''
snake_case_ :List[Any] = get_frameworks_table()
snake_case_ :str = Dataset.from_pandas(_lowercase )
snake_case_ :List[Any] = hf_hub_download(
"""huggingface/transformers-metadata""", """pipeline_tags.json""", repo_type="""dataset""", token=_lowercase )
snake_case_ :List[str] = Dataset.from_json(_lowercase )
snake_case_ :int = {
tags_dataset[i]["""model_class"""]: (tags_dataset[i]["""pipeline_tag"""], tags_dataset[i]["""auto_class"""])
for i in range(len(_lowercase ) )
}
snake_case_ :Optional[int] = update_pipeline_and_auto_class_table(_lowercase )
# Sort the model classes to avoid some nondeterministic updates to create false update commits.
snake_case_ :Tuple = sorted(table.keys() )
snake_case_ :Tuple = pd.DataFrame(
{
"""model_class""": model_classes,
"""pipeline_tag""": [table[m][0] for m in model_classes],
"""auto_class""": [table[m][1] for m in model_classes],
} )
snake_case_ :Union[str, Any] = Dataset.from_pandas(_lowercase )
with tempfile.TemporaryDirectory() as tmp_dir:
frameworks_dataset.to_json(os.path.join(_lowercase, """frameworks.json""" ) )
tags_dataset.to_json(os.path.join(_lowercase, """pipeline_tags.json""" ) )
if commit_sha is not None:
snake_case_ :Union[str, Any] = (
f"""Update with commit {commit_sha}\n\nSee: """
f"""https://github.com/huggingface/transformers/commit/{commit_sha}"""
)
else:
snake_case_ :List[Any] = """Update"""
upload_folder(
repo_id="""huggingface/transformers-metadata""", folder_path=_lowercase, repo_type="""dataset""", token=_lowercase, commit_message=_lowercase, )
def A_ ( ):
'''simple docstring'''
snake_case_ :List[Any] = {tag: cls for tag, _, cls in PIPELINE_TAGS_AND_AUTO_MODELS}
snake_case_ :Dict = transformers_module.pipelines.SUPPORTED_TASKS
snake_case_ :List[str] = []
for key in pipeline_tasks:
if key not in in_table:
snake_case_ :int = pipeline_tasks[key]["""pt"""]
if isinstance(_lowercase, (list, tuple) ):
snake_case_ :Any = model[0]
snake_case_ :str = model.__name__
if model not in in_table.values():
missing.append(_lowercase )
if len(_lowercase ) > 0:
snake_case_ :Optional[int] = """, """.join(_lowercase )
raise ValueError(
"""The following pipeline tags are not present in the `PIPELINE_TAGS_AND_AUTO_MODELS` constant inside """
f"""`utils/update_metadata.py`: {msg}. Please add them!""" )
if __name__ == "__main__":
__a = argparse.ArgumentParser()
parser.add_argument("--token", type=str, help="The token to use to push to the transformers-metadata dataset.")
parser.add_argument("--commit_sha", type=str, help="The sha of the commit going with this update.")
parser.add_argument("--check-only", action="store_true", help="Activate to just check all pipelines are present.")
__a = parser.parse_args()
if args.check_only:
check_pipeline_tags()
else:
update_metadata(args.token, args.commit_sha)
| 66 | 0 |
'''simple docstring'''
from cva import destroyAllWindows, imread, imshow, waitKey
def _SCREAMING_SNAKE_CASE ( UpperCamelCase ):
"""simple docstring"""
lowerCAmelCase__ , lowerCAmelCase__ : Union[str, Any] = img.shape[0], img.shape[1]
# converting each pixel's color to its negative
for i in range(UpperCamelCase ):
for j in range(UpperCamelCase ):
lowerCAmelCase__ : Union[str, Any] = [255, 255, 255] - img[i][j]
return img
if __name__ == "__main__":
# read original image
_lowerCAmelCase = imread('''image_data/lena.jpg''', 1)
# convert to its negative
_lowerCAmelCase = convert_to_negative(img)
# show result image
imshow('''negative of original image''', img)
waitKey(0)
destroyAllWindows()
| 37 |
"""simple docstring"""
import argparse
import glob
import logging
import os
from argparse import Namespace
from importlib import import_module
import numpy as np
import torch
from lightning_base import BaseTransformer, add_generic_args, generic_train
from seqeval.metrics import accuracy_score, fa_score, precision_score, recall_score
from torch.nn import CrossEntropyLoss
from torch.utils.data import DataLoader, TensorDataset
from utils_ner import TokenClassificationTask
__a = logging.getLogger(__name__)
class lowerCamelCase ( _lowerCAmelCase ):
'''simple docstring'''
_A : Union[str, Any] = """token-classification"""
def __init__( self: Any , snake_case: Tuple ) -> List[Any]:
if type(snake_case ) == dict:
snake_case_ :Optional[int] = Namespace(**snake_case )
snake_case_ :Optional[int] = import_module("""tasks""" )
try:
snake_case_ :Any = getattr(snake_case , hparams.task_type )
snake_case_ :TokenClassificationTask = token_classification_task_clazz()
except AttributeError:
raise ValueError(
f"""Task {hparams.task_type} needs to be defined as a TokenClassificationTask subclass in {module}. """
f"""Available tasks classes are: {TokenClassificationTask.__subclasses__()}""" )
snake_case_ :Any = self.token_classification_task.get_labels(hparams.labels )
snake_case_ :str = CrossEntropyLoss().ignore_index
super().__init__(snake_case , len(self.labels ) , self.mode )
def lowerCAmelCase_ ( self: Dict , **snake_case: List[Any] ) -> Any:
return self.model(**snake_case )
def lowerCAmelCase_ ( self: str , snake_case: Tuple , snake_case: List[Any] ) -> Optional[int]:
snake_case_ :List[str] = {"""input_ids""": batch[0], """attention_mask""": batch[1], """labels""": batch[3]}
if self.config.model_type != "distilbert":
snake_case_ :List[str] = (
batch[2] if self.config.model_type in ["""bert""", """xlnet"""] else None
) # XLM and RoBERTa don"t use token_type_ids
snake_case_ :Optional[Any] = self(**snake_case )
snake_case_ :List[str] = outputs[0]
# tensorboard_logs = {"loss": loss, "rate": self.lr_scheduler.get_last_lr()[-1]}
return {"loss": loss}
def lowerCAmelCase_ ( self: int ) -> Dict:
snake_case_ :List[Any] = self.hparams
for mode in ["train", "dev", "test"]:
snake_case_ :Optional[int] = self._feature_file(snake_case )
if os.path.exists(snake_case ) and not args.overwrite_cache:
logger.info("""Loading features from cached file %s""" , snake_case )
snake_case_ :Optional[int] = torch.load(snake_case )
else:
logger.info("""Creating features from dataset file at %s""" , args.data_dir )
snake_case_ :Optional[int] = self.token_classification_task.read_examples_from_file(args.data_dir , snake_case )
snake_case_ :Any = self.token_classification_task.convert_examples_to_features(
snake_case , self.labels , args.max_seq_length , self.tokenizer , cls_token_at_end=bool(self.config.model_type in ["""xlnet"""] ) , cls_token=self.tokenizer.cls_token , cls_token_segment_id=2 if self.config.model_type in ["""xlnet"""] else 0 , sep_token=self.tokenizer.sep_token , sep_token_extra=snake_case , pad_on_left=bool(self.config.model_type in ["""xlnet"""] ) , pad_token=self.tokenizer.pad_token_id , pad_token_segment_id=self.tokenizer.pad_token_type_id , pad_token_label_id=self.pad_token_label_id , )
logger.info("""Saving features into cached file %s""" , snake_case )
torch.save(snake_case , snake_case )
def lowerCAmelCase_ ( self: Optional[int] , snake_case: int , snake_case: int , snake_case: bool = False ) -> DataLoader:
snake_case_ :int = self._feature_file(snake_case )
logger.info("""Loading features from cached file %s""" , snake_case )
snake_case_ :str = torch.load(snake_case )
snake_case_ :Dict = torch.tensor([f.input_ids for f in features] , dtype=torch.long )
snake_case_ :str = torch.tensor([f.attention_mask for f in features] , dtype=torch.long )
if features[0].token_type_ids is not None:
snake_case_ :List[Any] = torch.tensor([f.token_type_ids for f in features] , dtype=torch.long )
else:
snake_case_ :List[str] = torch.tensor([0 for f in features] , dtype=torch.long )
# HACK(we will not use this anymore soon)
snake_case_ :Any = torch.tensor([f.label_ids for f in features] , dtype=torch.long )
return DataLoader(
TensorDataset(snake_case , snake_case , snake_case , snake_case ) , batch_size=snake_case )
def lowerCAmelCase_ ( self: List[str] , snake_case: Dict , snake_case: Union[str, Any] ) -> List[str]:
"""Compute validation""" ""
snake_case_ :List[str] = {"""input_ids""": batch[0], """attention_mask""": batch[1], """labels""": batch[3]}
if self.config.model_type != "distilbert":
snake_case_ :Dict = (
batch[2] if self.config.model_type in ["""bert""", """xlnet"""] else None
) # XLM and RoBERTa don"t use token_type_ids
snake_case_ :Dict = self(**snake_case )
snake_case_, snake_case_ :Dict = outputs[:2]
snake_case_ :Union[str, Any] = logits.detach().cpu().numpy()
snake_case_ :List[Any] = inputs["""labels"""].detach().cpu().numpy()
return {"val_loss": tmp_eval_loss.detach().cpu(), "pred": preds, "target": out_label_ids}
def lowerCAmelCase_ ( self: List[Any] , snake_case: int ) -> Tuple:
snake_case_ :Union[str, Any] = torch.stack([x["""val_loss"""] for x in outputs] ).mean()
snake_case_ :Tuple = np.concatenate([x["""pred"""] for x in outputs] , axis=0 )
snake_case_ :Tuple = np.argmax(snake_case , axis=2 )
snake_case_ :List[str] = np.concatenate([x["""target"""] for x in outputs] , axis=0 )
snake_case_ :Optional[Any] = dict(enumerate(self.labels ) )
snake_case_ :Dict = [[] for _ in range(out_label_ids.shape[0] )]
snake_case_ :Dict = [[] for _ in range(out_label_ids.shape[0] )]
for i in range(out_label_ids.shape[0] ):
for j in range(out_label_ids.shape[1] ):
if out_label_ids[i, j] != self.pad_token_label_id:
out_label_list[i].append(label_map[out_label_ids[i][j]] )
preds_list[i].append(label_map[preds[i][j]] )
snake_case_ :str = {
"""val_loss""": val_loss_mean,
"""accuracy_score""": accuracy_score(snake_case , snake_case ),
"""precision""": precision_score(snake_case , snake_case ),
"""recall""": recall_score(snake_case , snake_case ),
"""f1""": fa_score(snake_case , snake_case ),
}
snake_case_ :List[Any] = dict(results.items() )
snake_case_ :Union[str, Any] = results
return ret, preds_list, out_label_list
def lowerCAmelCase_ ( self: Optional[Any] , snake_case: Dict ) -> Optional[Any]:
# when stable
snake_case_, snake_case_, snake_case_ :Tuple = self._eval_end(snake_case )
snake_case_ :str = ret["""log"""]
return {"val_loss": logs["val_loss"], "log": logs, "progress_bar": logs}
def lowerCAmelCase_ ( self: Tuple , snake_case: Optional[int] ) -> Any:
# updating to test_epoch_end instead of deprecated test_end
snake_case_, snake_case_, snake_case_ :Any = self._eval_end(snake_case )
# Converting to the dict required by pl
# https://github.com/PyTorchLightning/pytorch-lightning/blob/master/\
# pytorch_lightning/trainer/logging.py#L139
snake_case_ :Optional[int] = ret["""log"""]
# `val_loss` is the key returned by `self._eval_end()` but actually refers to `test_loss`
return {"avg_test_loss": logs["val_loss"], "log": logs, "progress_bar": logs}
@staticmethod
def lowerCAmelCase_ ( snake_case: Any , snake_case: int ) -> Dict:
# Add NER specific options
BaseTransformer.add_model_specific_args(snake_case , snake_case )
parser.add_argument(
"""--task_type""" , default="""NER""" , type=snake_case , help="""Task type to fine tune in training (e.g. NER, POS, etc)""" )
parser.add_argument(
"""--max_seq_length""" , default=128 , type=snake_case , help=(
"""The maximum total input sequence length after tokenization. Sequences longer """
"""than this will be truncated, sequences shorter will be padded."""
) , )
parser.add_argument(
"""--labels""" , default="""""" , type=snake_case , help="""Path to a file containing all labels. If not specified, CoNLL-2003 labels are used.""" , )
parser.add_argument(
"""--gpus""" , default=0 , type=snake_case , help="""The number of GPUs allocated for this, it is by default 0 meaning none""" , )
parser.add_argument(
"""--overwrite_cache""" , action="""store_true""" , help="""Overwrite the cached training and evaluation sets""" )
return parser
if __name__ == "__main__":
__a = argparse.ArgumentParser()
add_generic_args(parser, os.getcwd())
__a = NERTransformer.add_model_specific_args(parser, os.getcwd())
__a = parser.parse_args()
__a = NERTransformer(args)
__a = generic_train(model, args)
if args.do_predict:
# See https://github.com/huggingface/transformers/issues/3159
# pl use this default format to create a checkpoint:
# https://github.com/PyTorchLightning/pytorch-lightning/blob/master\
# /pytorch_lightning/callbacks/model_checkpoint.py#L322
__a = sorted(glob.glob(os.path.join(args.output_dir, "checkpoint-epoch=*.ckpt"), recursive=True))
__a = model.load_from_checkpoint(checkpoints[-1])
trainer.test(model)
| 66 | 0 |
from collections import deque
from .hash_table import HashTable
class _SCREAMING_SNAKE_CASE ( _a ):
def __init__( self : Union[str, Any] , *__lowerCamelCase : Dict , **__lowerCamelCase : str ):
super().__init__(*__lowerCamelCase , **__lowerCamelCase )
def _A ( self : Optional[int] , __lowerCamelCase : Union[str, Any] , __lowerCamelCase : Union[str, Any] ):
UpperCamelCase :List[Any] = deque([] ) if self.values[key] is None else self.values[key]
self.values[key].appendleft(__lowerCamelCase )
UpperCamelCase :List[str] = self.values[key]
def _A ( self : List[Any] ):
return (
sum(self.charge_factor - len(__lowerCamelCase ) for slot in self.values )
/ self.size_table
* self.charge_factor
)
def _A ( self : Tuple , __lowerCamelCase : Dict , __lowerCamelCase : Dict=None ):
if not (
len(self.values[key] ) == self.charge_factor and self.values.count(__lowerCamelCase ) == 0
):
return key
return super()._collision_resolution(__lowerCamelCase , __lowerCamelCase )
| 38 |
"""simple docstring"""
from math import factorial
class lowerCamelCase :
'''simple docstring'''
def __init__( self: Optional[int] , snake_case: Dict , snake_case: int ) -> Tuple:
snake_case_ :List[Any] = real
if isinstance(snake_case , snake_case ):
snake_case_ :Tuple = [1] * rank
else:
snake_case_ :Optional[Any] = rank
def __repr__( self: List[str] ) -> Tuple:
return (
f"""{self.real}+"""
f"""{'+'.join(str(snake_case )+'E'+str(n+1 )for n,dual in enumerate(self.duals ) )}"""
)
def lowerCAmelCase_ ( self: Optional[int] ) -> Optional[int]:
snake_case_ :Any = self.duals.copy()
while cur[-1] == 0:
cur.pop(-1 )
return Dual(self.real , snake_case )
def __add__( self: Optional[int] , snake_case: Dict ) -> List[str]:
if not isinstance(snake_case , snake_case ):
return Dual(self.real + other , self.duals )
snake_case_ :List[Any] = self.duals.copy()
snake_case_ :Tuple = other.duals.copy()
if len(snake_case ) > len(snake_case ):
o_dual.extend([1] * (len(snake_case ) - len(snake_case )) )
elif len(snake_case ) < len(snake_case ):
s_dual.extend([1] * (len(snake_case ) - len(snake_case )) )
snake_case_ :Dict = []
for i in range(len(snake_case ) ):
new_duals.append(s_dual[i] + o_dual[i] )
return Dual(self.real + other.real , snake_case )
_A : str = __add__
def __sub__( self: Tuple , snake_case: Union[str, Any] ) -> Tuple:
return self + other * -1
def __mul__( self: str , snake_case: Tuple ) -> Optional[Any]:
if not isinstance(snake_case , snake_case ):
snake_case_ :Dict = []
for i in self.duals:
new_duals.append(i * other )
return Dual(self.real * other , snake_case )
snake_case_ :int = [0] * (len(self.duals ) + len(other.duals ) + 1)
for i, item in enumerate(self.duals ):
for j, jtem in enumerate(other.duals ):
new_duals[i + j + 1] += item * jtem
for k in range(len(self.duals ) ):
new_duals[k] += self.duals[k] * other.real
for index in range(len(other.duals ) ):
new_duals[index] += other.duals[index] * self.real
return Dual(self.real * other.real , snake_case )
_A : int = __mul__
def __truediv__( self: List[str] , snake_case: List[str] ) -> List[str]:
if not isinstance(snake_case , snake_case ):
snake_case_ :Optional[Any] = []
for i in self.duals:
new_duals.append(i / other )
return Dual(self.real / other , snake_case )
raise ValueError
def __floordiv__( self: int , snake_case: List[Any] ) -> Any:
if not isinstance(snake_case , snake_case ):
snake_case_ :Optional[int] = []
for i in self.duals:
new_duals.append(i // other )
return Dual(self.real // other , snake_case )
raise ValueError
def __pow__( self: Optional[Any] , snake_case: Optional[int] ) -> List[Any]:
if n < 0 or isinstance(snake_case , snake_case ):
raise ValueError("""power must be a positive integer""" )
if n == 0:
return 1
if n == 1:
return self
snake_case_ :str = self
for _ in range(n - 1 ):
x *= self
return x
def A_ ( _lowercase, _lowercase, _lowercase ):
'''simple docstring'''
if not callable(_lowercase ):
raise ValueError("""differentiate() requires a function as input for func""" )
if not isinstance(_lowercase, (float, int) ):
raise ValueError("""differentiate() requires a float as input for position""" )
if not isinstance(_lowercase, _lowercase ):
raise ValueError("""differentiate() requires an int as input for order""" )
snake_case_ :Optional[Any] = Dual(_lowercase, 1 )
snake_case_ :List[Any] = func(_lowercase )
if order == 0:
return result.real
return result.duals[order - 1] * factorial(_lowercase )
if __name__ == "__main__":
import doctest
doctest.testmod()
def A_ ( _lowercase ):
'''simple docstring'''
return y**2 * y**4
print(differentiate(f, 9, 2))
| 66 | 0 |
from typing import TYPE_CHECKING
from ...utils import (
OptionalDependencyNotAvailable,
_LazyModule,
is_tf_available,
is_torch_available,
is_vision_available,
)
_a = {
'''configuration_convnext''': ['''CONVNEXT_PRETRAINED_CONFIG_ARCHIVE_MAP''', '''ConvNextConfig''', '''ConvNextOnnxConfig''']
}
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_a = ['''ConvNextFeatureExtractor''']
_a = ['''ConvNextImageProcessor''']
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_a = [
'''CONVNEXT_PRETRAINED_MODEL_ARCHIVE_LIST''',
'''ConvNextForImageClassification''',
'''ConvNextModel''',
'''ConvNextPreTrainedModel''',
'''ConvNextBackbone''',
]
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
_a = [
'''TFConvNextForImageClassification''',
'''TFConvNextModel''',
'''TFConvNextPreTrainedModel''',
]
if TYPE_CHECKING:
from .configuration_convnext import CONVNEXT_PRETRAINED_CONFIG_ARCHIVE_MAP, ConvNextConfig, ConvNextOnnxConfig
try:
if not is_vision_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .feature_extraction_convnext import ConvNextFeatureExtractor
from .image_processing_convnext import ConvNextImageProcessor
try:
if not is_torch_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_convnext import (
CONVNEXT_PRETRAINED_MODEL_ARCHIVE_LIST,
ConvNextBackbone,
ConvNextForImageClassification,
ConvNextModel,
ConvNextPreTrainedModel,
)
try:
if not is_tf_available():
raise OptionalDependencyNotAvailable()
except OptionalDependencyNotAvailable:
pass
else:
from .modeling_tf_convnext import TFConvNextForImageClassification, TFConvNextModel, TFConvNextPreTrainedModel
else:
import sys
_a = _LazyModule(__name__, globals()['''__file__'''], _import_structure)
| 39 |
"""simple docstring"""
from __future__ import annotations
__a = 10
def A_ ( _lowercase ):
'''simple docstring'''
snake_case_ :Union[str, Any] = 1
snake_case_ :List[str] = max(_lowercase )
while placement <= max_digit:
# declare and initialize empty buckets
snake_case_ :list[list] = [[] for _ in range(_lowercase )]
# split list_of_ints between the buckets
for i in list_of_ints:
snake_case_ :Any = int((i / placement) % RADIX )
buckets[tmp].append(_lowercase )
# put each buckets' contents into list_of_ints
snake_case_ :Optional[Any] = 0
for b in range(_lowercase ):
for i in buckets[b]:
snake_case_ :Union[str, Any] = i
a += 1
# move to next
placement *= RADIX
return list_of_ints
if __name__ == "__main__":
import doctest
doctest.testmod()
| 66 | 0 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.