prune.py

# Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import functools
import os
from dataclasses import dataclass, field

import paddle
import paddle.nn.functional as F
from metric import MetricReport
from paddleslim.nas.ofa import OFA
from utils import preprocess_function, read_local_dataset

from paddlenlp.data import DataCollatorWithPadding
from paddlenlp.datasets import load_dataset
from paddlenlp.trainer import CompressionArguments, PdArgumentParser, Trainer
from paddlenlp.transformers import AutoModelForSequenceClassification, AutoTokenizer
from paddlenlp.utils.log import logger


# yapf: disable
@dataclass
class DataArguments:
    """
    Arguments pertaining to what data we are going to input our model for training and eval.
    Using `PdArgumentParser` we can turn this class
    into argparse arguments to be able to specify them on
    the command line.
    """

    dataset_dir: str = field(default=None, metadata={"help": "Local dataset directory should include train.txt, dev.txt and label.txt."})
    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."})


@dataclass
class ModelArguments:
    """
    Arguments pertaining to which model/config/tokenizer we are going to fine-tune from.
    """
    params_dir: str = field(default='./checkpoint/', metadata={"help": "The output directory where the model checkpoints are written."})
# yapf: enable


@paddle.no_grad()
def custom_evaluate(self, model, data_loader):
    metric = MetricReport()
    model.eval()
    metric.reset()
    for batch in data_loader:
        logits = model(batch["input_ids"], batch["token_type_ids"])
        # Supports paddleslim.nas.ofa.OFA model and nn.layer model.
        if isinstance(model, OFA):
            logits = logits[0]
        probs = F.sigmoid(logits)
        metric.update(probs, batch["labels"])

    micro_f1_score, macro_f1_score = metric.accumulate()
    logger.info("micro f1 score: %.5f, macro f1 score: %.5f" % (micro_f1_score, macro_f1_score))
    model.train()
    return macro_f1_score


def main():
    parser = PdArgumentParser((ModelArguments, DataArguments, CompressionArguments))
    model_args, data_args, compression_args = parser.parse_args_into_dataclasses()
    paddle.set_device(compression_args.device)
    compression_args.strategy = "dynabert"
    # Log model and data config
    compression_args.print_config(model_args, "Model")
    compression_args.print_config(data_args, "Data")

    label_list = {}
    label_path = os.path.join(data_args.dataset_dir, "label.txt")
    train_path = os.path.join(data_args.dataset_dir, "train.txt")
    dev_path = os.path.join(data_args.dataset_dir, "dev.txt")
    with open(label_path, "r", encoding="utf-8") as f:
        for i, line in enumerate(f):
            l = line.strip()
            label_list[l] = i

    train_ds = load_dataset(read_local_dataset, path=train_path, label_list=label_list, lazy=False)
    dev_ds = load_dataset(read_local_dataset, path=dev_path, label_list=label_list, lazy=False)

    model = AutoModelForSequenceClassification.from_pretrained(model_args.params_dir)
    tokenizer = AutoTokenizer.from_pretrained(model_args.params_dir)

    trans_func = functools.partial(
        preprocess_function, tokenizer=tokenizer, max_seq_length=data_args.max_seq_length, label_nums=len(label_list)
    )
    train_dataset = train_ds.map(trans_func)
    dev_dataset = dev_ds.map(trans_func)

    # Define data collector， criterion
    data_collator = DataCollatorWithPadding(tokenizer)
    criterion = paddle.nn.BCEWithLogitsLoss()

    trainer = Trainer(
        model=model,
        args=compression_args,
        data_collator=data_collator,
        train_dataset=train_dataset,
        eval_dataset=dev_dataset,
        criterion=criterion,
    )  # Strategy`dynabert` needs arguments `criterion`

    compression_args.print_config()

    trainer.compress(custom_evaluate=custom_evaluate)


if __name__ == "__main__":
    main()