engine_pretrain.py

# Copyright (c) 2023 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.

# --------------------------------------------------------
# References:
# MAE: https://github.com/facebookresearch/mae
# --------------------------------------------------------
import math
import sys
from typing import Iterable

import paddle

import util.misc as misc
import util.lr_sched as lr_sched


def train_one_epoch(model,
                    data_loader,
                    optimizer,
                    device,
                    epoch,
                    loss_scaler,
                    log_writer=None,
                    args=None):
    model.train()
    metric_logger = misc.MetricLogger(delimiter="  ")
    metric_logger.add_meter(
        'lr', misc.SmoothedValue(
            window_size=1, fmt='{value:.6f}'))
    metric_logger.add_meter(
        'loss', misc.SmoothedValue(
            window_size=1, fmt='{value:.4f}'))
    header = 'Epoch: [{}]'.format(epoch)

    accum_iter = args.accum_iter

    optimizer.clear_grad()

    if log_writer is not None:
        print('log_dir: {}'.format(log_writer.kwargs['log_dir']))

    for data_iter_step, (samples, _) in enumerate(
            metric_logger.log_every(data_loader, args.print_freq, header)):

        global_iter_step = data_iter_step + len(data_loader) * epoch
        if args.max_train_step is not None and global_iter_step >= args.max_train_step:
            print(
                f'step({global_iter_step}) >= max_train_step({args.max_train_step}), training stops early. This function is only used for debugging.'
            )
            exit(0)

        # we use a per iteration (instead of per epoch) lr scheduler
        if data_iter_step % accum_iter == 0:
            lr_sched.adjust_learning_rate(
                optimizer, data_iter_step / len(data_loader) + epoch, args)

        with paddle.amp.auto_cast():
            loss, _, _ = model(samples, mask_ratio=args.mask_ratio)

        loss_value = loss.item()

        if not math.isfinite(loss_value):
            print("Loss is {}, stopping training".format(loss_value))
            sys.exit(1)

        loss /= accum_iter
        loss_scaler(
            loss,
            optimizer,
            parameters=model.parameters(),
            update_grad=(data_iter_step + 1) % accum_iter == 0)
        if (data_iter_step + 1) % accum_iter == 0:
            optimizer.clear_grad()

        paddle.device.cuda.synchronize()

        metric_logger.update(loss=loss_value)

        lr = optimizer._param_groups[0]["lr"]
        metric_logger.update(lr=lr)

        loss_value_reduce = misc.all_reduce_mean(loss_value)
        if log_writer is not None and (data_iter_step + 1) % accum_iter == 0:
            """ We use epoch_1000x as the x-axis in tensorboard.
            This calibrates different curves when batch size changes.
            """
            epoch_1000x = int(
                (data_iter_step / len(data_loader) + epoch) * 1000)
            log_writer.add_scalar('train_loss', loss_value_reduce, epoch_1000x)
            log_writer.add_scalar('lr', lr, epoch_1000x)

    # gather the stats from all processes
    metric_logger.synchronize_between_processes()
    print("Averaged stats:", metric_logger)
    return {k: meter.global_avg for k, meter in metric_logger.meters.items()}