run_glue_from_dcl.py

# coding=utf-8

import dataclasses
import logging
import os
import sys
from dataclasses import dataclass, field
from typing import Callable, Dict, Optional
import torch
import numpy as np
import yaml
from transformers import AutoConfig, AutoModelForSequenceClassification, AutoTokenizer, EvalPrediction, GlueDataset
from transformers import GlueDataTrainingArguments as DataTrainingArguments
from transformers import (
    HfArgumentParser,
    Trainer,
    TrainingArguments,
    glue_compute_metrics,
    glue_output_modes,
    glue_tasks_num_labels,
    set_seed,
    RobertaTokenizer,
    RobertaForMaskedLM,
    BertForMaskedLM,
    BertTokenizer,
)
from models.mlp_head import MLPHead
from models.bert_base_network import BertNet
from da_utils import GlueDaDataset


logger = logging.getLogger(__name__)


@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"}
    )
    cache_dir: Optional[str] = field(
        default=None, metadata={"help": "Where do you want to store the pretrained models downloaded from s3"}
    )
def prepare_bert(model_path):
    if model_path == 'bert-base-uncased':
        model_class = BertForMaskedLM
        tokenizer_class = BertTokenizer
        pretrained_weights = 'bert-base-uncased'
    elif model_path == 'roberta-base':
        model_class = RobertaForMaskedLM
        tokenizer_class = RobertaTokenizer
        pretrained_weights = 'roberta-base'


    # 载入预训练的tokenizer和model
    tokenizer = tokenizer_class.from_pretrained(pretrained_weights)
    model = model_class.from_pretrained(pretrained_weights)
    # model1 = RobertaModel.from_pretrained(pretrained_weights)
    # id = tokenizer.encode("Here is some text to encode", add_special_tokens=True)
    # input_ids = torch.tensor([tokenizer.encode("Here is some text to encode", add_special_tokens=True)])
    # with torch.no_grad():
    #     last_hidden_states = model(input_ids)[0]  # Models outputs are now tuples
    return tokenizer,model


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.
    byol_config = yaml.load(open("./config/config.yaml", "r"), Loader=yaml.FullLoader)
    parser = HfArgumentParser((ModelArguments, DataTrainingArguments, TrainingArguments))

    if len(sys.argv) == 2 and sys.argv[1].endswith(".json"):
        # If we pass only one argument to the script and it's the path to a json file,
        # let's parse it to get our arguments.
        model_args, data_args, training_args = parser.parse_json_file(json_file=os.path.abspath(sys.argv[1]))
    else:
        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 if training_args.local_rank in [-1, 0] else logging.WARN,
    )
    logger.warning(
        "Process rank: %s, device: %s, n_gpu: %s, distributed training: %s, 16-bits training: %s",
        training_args.local_rank,
        training_args.device,
        training_args.n_gpu,
        bool(training_args.local_rank != -1),
        training_args.fp16,
    )
    logger.info("Training/evaluation parameters %s", training_args)

    # Set seed
    set_seed(training_args.seed)

    try:
        num_labels = glue_tasks_num_labels[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,
    )
    model = AutoModelForSequenceClassification.from_pretrained(
        # os.path.join(model_args.model_name_or_path,'checkpoint-60000'),
        model_args.model_name_or_path,
        from_tf=bool(".ckpt" in model_args.model_name_or_path),
        config=config,
        cache_dir=model_args.cache_dir,
    )



    # load contrast-trained parameters
    _, robertaformlm = prepare_bert(model_path=model_args.model_name_or_path)
    robertaformlm.resize_token_embeddings(len(tokenizer))
    online_network = BertNet(bert=robertaformlm,norm_type='power',model_name=model_args.model_name_or_path, **byol_config['network']) # data_args.task_name.upper()
    load_params = torch.load(os.path.join('/home/chenxiang/pretrained_model/PyTorch-BYOL/new_result/bert_power_pretrain_glue',data_args.task_name.upper(),'pytorch_model.bin'))
    if 'online_network_state_dict' in load_params:
       online_network.load_state_dict(load_params['online_network_state_dict'])
       model.bert.load_state_dict(online_network.roberta.state_dict())
       del online_network, load_params,robertaformlm
       print("Parameters successfully loaded.")


    # Get datasets
    train_dataset = (
        GlueDataset(data_args, tokenizer=tokenizer, cache_dir=model_args.cache_dir) if training_args.do_train else None
    )
    eval_dataset = (
        GlueDataset(data_args, tokenizer=tokenizer, mode="dev", cache_dir=model_args.cache_dir)
        if training_args.do_eval
        else None
    )
    test_dataset = (
        GlueDataset(data_args, tokenizer=tokenizer, mode="test", cache_dir=model_args.cache_dir)
        if training_args.do_predict
        else None
    )

    def build_compute_metrics_fn(task_name: str) -> Callable[[EvalPrediction], Dict]:
        def compute_metrics_fn(p: EvalPrediction):
            if output_mode == "classification":
                preds = np.argmax(p.predictions, axis=1)
            elif output_mode == "regression":
                preds = np.squeeze(p.predictions)
            return glue_compute_metrics(task_name, preds, p.label_ids)

        return compute_metrics_fn

    dataset_num = len(train_dataset)
    training_args.save_steps = dataset_num//training_args.per_device_eval_batch_size//2
    training_args.eval_steps = training_args.save_steps
    training_args.logging_steps = training_args.save_steps
    print('save_steps:', training_args.save_steps)
    print('eval_steps:', training_args.eval_steps)
    print('logging_steps:', training_args.logging_steps)

    # Initialize our Trainer
    trainer = Trainer(
        model=model,
        args=training_args,
        train_dataset=train_dataset,
        eval_dataset=eval_dataset,
        is_earlystop=False,
        compute_metrics=build_compute_metrics_fn(data_args.task_name),
    )

    # Training
    if training_args.do_train:
        trainer.train(
            # model_path=None
            # model_path='roberta_contrast_finetune/MRPC/checkpoint-460_best92.46'
            model_args.model_name_or_path if os.path.isdir(model_args.model_name_or_path) else None
        )
        trainer.save_model()
        # For convenience, we also re-save the tokenizer to the same directory,
        # so that you can share your model easily on huggingface.co/models =)
        if trainer.is_world_master():
            tokenizer.save_pretrained(training_args.output_dir)

    # Evaluation
    eval_results = {}
    if training_args.do_eval:
        logger.info("*** Evaluate ***")

        # Loop to handle MNLI double evaluation (matched, mis-matched)
        eval_datasets = [eval_dataset]
        if data_args.task_name == "mnli":
            mnli_mm_data_args = dataclasses.replace(data_args, task_name="mnli-mm")
            eval_datasets.append(
                GlueDaDataset(mnli_mm_data_args, tokenizer=tokenizer, mode="dev", cache_dir=model_args.cache_dir)
            )

        for eval_dataset in eval_datasets:
            trainer.compute_metrics = build_compute_metrics_fn(eval_dataset.args.task_name)
            eval_result = trainer.evaluate(eval_dataset=eval_dataset)

            output_eval_file = os.path.join(
                training_args.output_dir, f"eval_results_{eval_dataset.args.task_name}.txt"
            )
            if trainer.is_world_master():
                with open(output_eval_file, "w") as writer:
                    logger.info("***** Eval results {} *****".format(eval_dataset.args.task_name))
                    for key, value in eval_result.items():
                        logger.info("  %s = %s", key, value)
                        writer.write("%s = %s\n" % (key, value))

            eval_results.update(eval_result)

    if training_args.do_predict:
        logging.info("*** Test ***")
        test_datasets = [test_dataset]
        if data_args.task_name == "mnli":
            mnli_mm_data_args = dataclasses.replace(data_args, task_name="mnli-mm")
            test_datasets.append(
                GlueDataset(mnli_mm_data_args, tokenizer=tokenizer, mode="test", cache_dir=model_args.cache_dir)
            )

        for test_dataset in test_datasets:
            predictions = trainer.predict(test_dataset=test_dataset).predictions
            if output_mode == "classification":
                predictions = np.argmax(predictions, axis=1)

            output_test_file = os.path.join(
                training_args.output_dir, f"test_results_{test_dataset.args.task_name}.txt"
            )
            if trainer.is_world_master():
                with open(output_test_file, "w") as writer:
                    logger.info("***** Test results {} *****".format(test_dataset.args.task_name))
                    writer.write("index\tprediction\n")
                    for index, item in enumerate(predictions):
                        if output_mode == "regression":
                            writer.write("%d\t%3.3f\n" % (index, item))
                        else:
                            item = test_dataset.get_labels()[item]
                            writer.write("%d\t%s\n" % (index, item))
    return eval_results


def _mp_fn(index):
    # For xla_spawn (TPUs)
    main()


if __name__ == "__main__":
    main()