mirror of
https://github.com/huggingface/transformers.git
synced 2025-07-03 21:00:08 +06:00

* Created using Colaboratory * [examples] reorganize files * remove run_tpu_glue.py as superseded by TPU support in Trainer * Bugfix: int, not tuple * move files around
230 lines
7.3 KiB
Python
230 lines
7.3 KiB
Python
# coding=utf-8
|
|
""" Fine-tuning the library models for sequence classification."""
|
|
|
|
|
|
import logging
|
|
import os
|
|
from dataclasses import dataclass, field
|
|
from enum import Enum
|
|
from typing import Dict, Optional
|
|
|
|
import numpy as np
|
|
import tensorflow_datasets as tfds
|
|
|
|
from transformers import (
|
|
AutoConfig,
|
|
AutoTokenizer,
|
|
EvalPrediction,
|
|
HfArgumentParser,
|
|
PreTrainedTokenizer,
|
|
TFAutoModelForSequenceClassification,
|
|
TFTrainer,
|
|
TFTrainingArguments,
|
|
glue_compute_metrics,
|
|
glue_convert_examples_to_features,
|
|
glue_output_modes,
|
|
glue_processors,
|
|
glue_tasks_num_labels,
|
|
)
|
|
|
|
|
|
class Split(Enum):
|
|
train = "train"
|
|
dev = "validation"
|
|
test = "test"
|
|
|
|
|
|
def get_tfds(
|
|
task_name: str, tokenizer: PreTrainedTokenizer, max_seq_length: Optional[int] = None, mode: Split = Split.train
|
|
):
|
|
if task_name == "mnli-mm" and mode == Split.dev:
|
|
tfds_name = "mnli_mismatched"
|
|
elif task_name == "mnli-mm" and mode == Split.train:
|
|
tfds_name = "mnli"
|
|
elif task_name == "mnli" and mode == Split.dev:
|
|
tfds_name = "mnli_matched"
|
|
elif task_name == "sst-2":
|
|
tfds_name = "sst2"
|
|
elif task_name == "sts-b":
|
|
tfds_name = "stsb"
|
|
else:
|
|
tfds_name = task_name
|
|
|
|
ds = tfds.load("glue/" + tfds_name, split=mode.value)
|
|
|
|
return glue_convert_examples_to_features(ds, tokenizer, max_seq_length, task_name)
|
|
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
|
|
@dataclass
|
|
class GlueDataTrainingArguments:
|
|
"""
|
|
Arguments pertaining to what data we are going to input our model for training and eval.
|
|
|
|
Using `HfArgumentParser` we can turn this class
|
|
into argparse arguments to be able to specify them on
|
|
the command line.
|
|
"""
|
|
|
|
task_name: str = field(metadata={"help": "The name of the task to train on: " + ", ".join(glue_processors.keys())})
|
|
max_seq_length: int = field(
|
|
default=128,
|
|
metadata={
|
|
"help": "The maximum total input sequence length after tokenization. Sequences longer "
|
|
"than this will be truncated, sequences shorter will be padded."
|
|
},
|
|
)
|
|
overwrite_cache: bool = field(
|
|
default=False, metadata={"help": "Overwrite the cached training and evaluation sets"}
|
|
)
|
|
|
|
def __post_init__(self):
|
|
self.task_name = self.task_name.lower()
|
|
|
|
|
|
@dataclass
|
|
class ModelArguments:
|
|
"""
|
|
Arguments pertaining to which model/config/tokenizer we are going to fine-tune from.
|
|
"""
|
|
|
|
model_name_or_path: str = field(
|
|
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models"}
|
|
)
|
|
config_name: Optional[str] = field(
|
|
default=None, metadata={"help": "Pretrained config name or path if not the same as model_name"}
|
|
)
|
|
tokenizer_name: Optional[str] = field(
|
|
default=None, metadata={"help": "Pretrained tokenizer name or path if not the same as model_name"}
|
|
)
|
|
use_fast: bool = field(default=False, metadata={"help": "Set this flag to use fast tokenization."})
|
|
# If you want to tweak more attributes on your tokenizer, you should do it in a distinct script,
|
|
# or just modify its tokenizer_config.json.
|
|
cache_dir: Optional[str] = field(
|
|
default=None, metadata={"help": "Where do you want to store the pretrained models downloaded from s3"}
|
|
)
|
|
|
|
|
|
def main():
|
|
# See all possible arguments in src/transformers/training_args.py
|
|
# or by passing the --help flag to this script.
|
|
# We now keep distinct sets of args, for a cleaner separation of concerns.
|
|
parser = HfArgumentParser((ModelArguments, GlueDataTrainingArguments, TFTrainingArguments))
|
|
model_args, data_args, training_args = parser.parse_args_into_dataclasses()
|
|
|
|
if (
|
|
os.path.exists(training_args.output_dir)
|
|
and os.listdir(training_args.output_dir)
|
|
and training_args.do_train
|
|
and not training_args.overwrite_output_dir
|
|
):
|
|
raise ValueError(
|
|
f"Output directory ({training_args.output_dir}) already exists and is not empty. Use --overwrite_output_dir to overcome."
|
|
)
|
|
|
|
# Setup logging
|
|
logging.basicConfig(
|
|
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
|
|
datefmt="%m/%d/%Y %H:%M:%S",
|
|
level=logging.INFO,
|
|
)
|
|
logger.info(
|
|
"n_gpu: %s, distributed training: %s, 16-bits training: %s",
|
|
training_args.n_gpu,
|
|
bool(training_args.n_gpu > 1),
|
|
training_args.fp16,
|
|
)
|
|
logger.info("Training/evaluation parameters %s", training_args)
|
|
|
|
try:
|
|
num_labels = glue_tasks_num_labels["mnli" if data_args.task_name == "mnli-mm" else data_args.task_name]
|
|
output_mode = glue_output_modes[data_args.task_name]
|
|
except KeyError:
|
|
raise ValueError("Task not found: %s" % (data_args.task_name))
|
|
|
|
# Load pretrained model and tokenizer
|
|
#
|
|
# Distributed training:
|
|
# The .from_pretrained methods guarantee that only one local process can concurrently
|
|
# download model & vocab.
|
|
|
|
config = AutoConfig.from_pretrained(
|
|
model_args.config_name if model_args.config_name else model_args.model_name_or_path,
|
|
num_labels=num_labels,
|
|
finetuning_task=data_args.task_name,
|
|
cache_dir=model_args.cache_dir,
|
|
)
|
|
tokenizer = AutoTokenizer.from_pretrained(
|
|
model_args.tokenizer_name if model_args.tokenizer_name else model_args.model_name_or_path,
|
|
cache_dir=model_args.cache_dir,
|
|
)
|
|
|
|
with training_args.strategy.scope():
|
|
model = TFAutoModelForSequenceClassification.from_pretrained(
|
|
model_args.model_name_or_path,
|
|
from_pt=bool(".bin" in model_args.model_name_or_path),
|
|
config=config,
|
|
cache_dir=model_args.cache_dir,
|
|
)
|
|
|
|
# Get datasets
|
|
train_dataset = (
|
|
get_tfds(task_name=data_args.task_name, tokenizer=tokenizer, max_seq_length=data_args.max_seq_length)
|
|
if training_args.do_train
|
|
else None
|
|
)
|
|
eval_dataset = (
|
|
get_tfds(
|
|
task_name=data_args.task_name, tokenizer=tokenizer, max_seq_length=data_args.max_seq_length, mode=Split.dev
|
|
)
|
|
if training_args.do_eval
|
|
else None
|
|
)
|
|
|
|
def compute_metrics(p: EvalPrediction) -> Dict:
|
|
if output_mode == "classification":
|
|
preds = np.argmax(p.predictions, axis=1)
|
|
elif output_mode == "regression":
|
|
preds = np.squeeze(p.predictions)
|
|
return glue_compute_metrics(data_args.task_name, preds, p.label_ids)
|
|
|
|
# Initialize our Trainer
|
|
trainer = TFTrainer(
|
|
model=model,
|
|
args=training_args,
|
|
train_dataset=train_dataset,
|
|
eval_dataset=eval_dataset,
|
|
compute_metrics=compute_metrics,
|
|
)
|
|
|
|
# Training
|
|
if training_args.do_train:
|
|
trainer.train()
|
|
trainer.save_model()
|
|
tokenizer.save_pretrained(training_args.output_dir)
|
|
|
|
# Evaluation
|
|
results = {}
|
|
if training_args.do_eval:
|
|
logger.info("*** Evaluate ***")
|
|
|
|
result = trainer.evaluate()
|
|
output_eval_file = os.path.join(training_args.output_dir, "eval_results.txt")
|
|
|
|
with open(output_eval_file, "w") as writer:
|
|
logger.info("***** Eval results *****")
|
|
|
|
for key, value in result.items():
|
|
logger.info(" %s = %s", key, value)
|
|
writer.write("%s = %s\n" % (key, value))
|
|
|
|
results.update(result)
|
|
|
|
return results
|
|
|
|
|
|
if __name__ == "__main__":
|
|
main()
|