engine/pretrain_amp.py

# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.

# This source code is licensed under the license found in the
# LICENSE file in the root directory of this source tree.
# --------------------------------------------------------
# References:
# DeiT: https://github.com/facebookresearch/deit
# BEiT: https://github.com/microsoft/unilm/tree/master/beit
# --------------------------------------------------------
import math
import sys
from typing import Iterable

import torch

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


def train_one_epoch(
    model,
    data_loader,
    optimizer,
    device,
    epoch: int,
    loss_scaler,
    log_writer=None,
    args=None,
):
    model.train(True)
    metric_logger = misc.MetricLogger(delimiter="  ")
    metric_logger.add_meter("lr", misc.SmoothedValue(window_size=1, fmt="{value:.6f}"))
    header = "Epoch: [{}]".format(epoch)
    print_freq = 20

    optimizer.zero_grad()

    if log_writer is not None:
        print("log_dir: {}".format(log_writer.log_dir))

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

        # we use a per iteration (instead of per epoch) lr scheduler
        samples = samples.to(device, non_blocking=True)
        lr_sched.adjust_learning_rate(
            optimizer, data_iter_step / len(data_loader) + epoch, args
        )

        with torch.cuda.amp.autocast():
            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_scaler(loss, optimizer, parameters=model.parameters(), clip_grad=1.0)
        optimizer.zero_grad()
        torch.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:
            """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()}