train.py 5.4 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
# copyright (c) 2020 PaddlePaddle Authors. All Rights Reserve.
#
# 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.

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import argparse
import os
import sys
__dir__ = os.path.dirname(os.path.abspath(__file__))
sys.path.append(__dir__)
sys.path.append(os.path.abspath(os.path.join(__dir__, '../../')))

from sys import version_info

import paddle
from paddle.distributed import fleet

from ppcls.data import Reader
from ppcls.utils.config import get_config
from ppcls.utils import logger
from tools.static import program
from program import save_model


def parse_args():
    parser = argparse.ArgumentParser("PaddleClas train script")
    parser.add_argument(
        '-c',
        '--config',
        type=str,
        default='configs/ResNet/ResNet50.yaml',
        help='config file path')
    parser.add_argument(
        '--vdl_dir',
        type=str,
        default=None,
        help='VisualDL logging directory for image.')
    parser.add_argument(
        '-o',
        '--override',
        action='append',
        default=[],
        help='config options to be overridden')
    args = parser.parse_args()
    return args


def main(args):
    fleet.init(is_collective=True)

    config = get_config(args.config, overrides=args.override, show=True)
    # assign the place
66 67 68 69 70 71
    use_gpu = config.get("use_gpu", False)
    use_xpu = config.get("use_xpu", False)
    assert (use_gpu or use_xpu) is True, "gpu or xpu must be true in static mode!"
    assert (use_gpu and use_xpu) is not True, "gpu and xpu can not be true in the same time in static mode!"

    place = paddle.set_device('gpu' if use_gpu else 'xpu')
72 73 74 75 76 77 78 79 80

    # startup_prog is used to do some parameter init work,
    # and train prog is used to hold the network
    startup_prog = paddle.static.Program()
    train_prog = paddle.static.Program()

    best_top1_acc = 0.0  # best top1 acc record

    train_fetchs, lr_scheduler, train_feeds = program.build(
81
        config, train_prog, startup_prog, is_train=True, is_distributed=config.get("is_distributed", True))
82 83 84 85

    if config.validate:
        valid_prog = paddle.static.Program()
        valid_fetchs, _, valid_feeds = program.build(
86
            config, valid_prog, startup_prog, is_train=False, is_distributed=config.get("is_distributed", True))
87 88 89 90 91 92 93 94 95 96 97
        # clone to prune some content which is irrelevant in valid_prog
        valid_prog = valid_prog.clone(for_test=True)

    # create the "Executor" with the statement of which place
    exe = paddle.static.Executor(place)
    # Parameter initialization
    exe.run(startup_prog)

    # load model from 1. checkpoint to resume training, 2. pretrained model to finetune
    train_dataloader = Reader(config, 'train', places=place)()

98
    if config.validate and paddle.distributed.get_rank() == 0:
99
        valid_dataloader = Reader(config, 'valid', places=place)()
100 101 102 103
        if use_xpu:
            compiled_valid_prog = valid_prog
        else:
            compiled_valid_prog = program.compile(config, valid_prog)
104 105 106 107 108 109 110 111 112 113 114 115 116

    vdl_writer = None
    if args.vdl_dir:
        if version_info.major == 2:
            logger.info(
                "visualdl is just supported for python3, so it is disabled in python2..."
            )
        else:
            from visualdl import LogWriter
            vdl_writer = LogWriter(args.vdl_dir)

    for epoch_id in range(config.epochs):
        # 1. train with train dataset
117 118 119 120
        program.run(train_dataloader, exe, train_prog, train_feeds,
                    train_fetchs, epoch_id, 'train', config, vdl_writer,
                    lr_scheduler)
        if paddle.distributed.get_rank() == 0:
121 122 123
            # 2. validate with validate dataset
            if config.validate and epoch_id % config.valid_interval == 0:
                top1_acc = program.run(valid_dataloader, exe,
124 125
                                       compiled_valid_prog, valid_feeds,
                                       valid_fetchs, epoch_id, 'valid', config)
126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147
                if top1_acc > best_top1_acc:
                    best_top1_acc = top1_acc
                    message = "The best top1 acc {:.5f}, in epoch: {:d}".format(
                        best_top1_acc, epoch_id)
                    logger.info("{:s}".format(logger.coloring(message, "RED")))
                    if epoch_id % config.save_interval == 0:

                        model_path = os.path.join(config.model_save_dir,
                                                  config.ARCHITECTURE["name"])
                        save_model(train_prog, model_path, "best_model")

            # 3. save the persistable model
            if epoch_id % config.save_interval == 0:
                model_path = os.path.join(config.model_save_dir,
                                          config.ARCHITECTURE["name"])
                save_model(train_prog, model_path, epoch_id)


if __name__ == '__main__':
    paddle.enable_static()
    args = parse_args()
    main(args)