train.py

# -*- coding: UTF-8 -*-
#   Copyright (c) 2019 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 os
import sys
import math
import time
import random
import argparse
import multiprocessing

import numpy as np
import paddle
import paddle.fluid as fluid

np.set_printoptions(threshold=np.inf)
import reader
import utils
from sequence_labeling import lex_net, Chunk_eval
#from eval import test_process

# the function to train model
def do_train(args):

    dataset = reader.Dataset(args)
    if args.use_cuda: 
        place = fluid.CUDAPlace(fluid.dygraph.parallel.Env().dev_id) \
        if args.use_data_parallel else fluid.CUDAPlace(0)
    else:
        place = fluid.CPUPlace()

    with fluid.dygraph.guard(place):
        if args.use_data_parallel:
            strategy = fluid.dygraph.parallel.prepare_context()
        #fluid.default_startup_program().random_seed = 102
        #fluid.default_main_program().random_seed = 102
        #np.random.seed(102)
        #random.seed(102)
        train_loader = reader.create_dataloader(
            args,
            file_name=args.train_data,
            place=place,
            model='lac',
            reader=dataset)
        if args.use_data_parallel:
            train_loader = fluid.contrib.reader.distributed_batch_reader(
                train_loader)

        test_loader = reader.create_dataloader(
            args,
            file_name=args.test_data,
            place=place,
            model='lac',
            reader=dataset,
            mode='test')
        model = lex_net(args, dataset.vocab_size, dataset.num_labels)
        if args.use_data_parallel:
            model = fluid.dygraph.parallel.DataParallel(model, strategy)
        optimizer = fluid.optimizer.AdamOptimizer(learning_rate=args.base_learning_rate,
                                                parameter_list=model.parameters())
        chunk_eval = Chunk_eval(int(math.ceil((dataset.num_labels - 1) / 2.0)), "IOB")
        num_train_examples = dataset.get_num_examples(args.train_data)
        max_train_steps = args.epoch * num_train_examples // args.batch_size
        print("Num train examples: %d" % num_train_examples)
        print("Max train steps: %d" % max_train_steps)

        step = 0
        print_start_time = time.time()
        chunk_evaluator = fluid.metrics.ChunkEvaluator()
        chunk_evaluator.reset()

        def test_process(reader, chunk_evaluator):
            model.eval()
            chunk_evaluator.reset()

            start_time = time.time()
            for batch in reader():
                words, targets, length = batch
                crf_decode = model(words, length=length)
                (precision, recall, f1_score, num_infer_chunks, num_label_chunks,
                    num_correct_chunks) = chunk_eval(
                        input=crf_decode,
                        label=targets,
                        seq_length=length)
                chunk_evaluator.update(num_infer_chunks.numpy(), num_label_chunks.numpy(), num_correct_chunks.numpy())
            
            precision, recall, f1 = chunk_evaluator.eval()
            end_time = time.time()
            print("[test] P: %.5f, R: %.5f, F1: %.5f, elapsed time: %.3f s" %
                (precision, recall, f1, end_time - start_time))
            model.train()

        for epoch_id in range(args.epoch):
            for batch in train_loader():
                words, targets, length = batch

                start_time = time.time()
                avg_cost, crf_decode = model(words, targets, length)
                if args.use_data_parallel:
                    avg_cost = model.scale_loss(avg_cost)
                    avg_cost.backward()
                    model.apply_collective_grads()
                else:
                    avg_cost.backward()
                optimizer.minimize(avg_cost)
                model.clear_gradients()
                end_time = time.time()

                if step % args.print_steps == 0:
                    (precision, recall, f1_score, num_infer_chunks, num_label_chunks,
                        num_correct_chunks) = chunk_eval(
                        input=crf_decode,
                        label=targets,
                        seq_length=length)
                    outputs = [avg_cost, precision, recall, f1_score]
                    avg_cost, precision, recall, f1_score = [np.mean(x.numpy()) for x in outputs]

                    print("[train] step = %d, loss = %.5f, P: %.5f, R: %.5f, F1: %.5f, elapsed time %.5f" % (
                        step, avg_cost, precision, recall, f1_score, end_time - start_time))

                if step % args.validation_steps == 0:
                    test_process(test_loader, chunk_evaluator)

                # save checkpoints
                if step % args.save_steps == 0 and step != 0:
                    save_path = os.path.join(args.model_save_dir, "step_" + str(step))
                    paddle.fluid.save_dygraph(model.state_dict(), save_path)
                step += 1

        

if __name__ == "__main__":
    # 参数控制可以根据需求使用argparse，yaml或者json
    # 对NLP任务推荐使用PALM下定义的configure，可以统一argparse，yaml或者json格式的配置文件。

    parser = argparse.ArgumentParser(__doc__)
    utils.load_yaml(parser, 'conf/args.yaml')

    args = parser.parse_args()

    print(args)

    do_train(args)