Open-Sora/scripts/train.py

import os
from copy import deepcopy
from datetime import timedelta
from pprint import pformat

import torch
import torch.distributed as dist
from colossalai.booster import Booster
from colossalai.cluster import DistCoordinator
from colossalai.nn.optimizer import HybridAdam
from colossalai.utils import get_current_device, set_seed
from tqdm import tqdm

import wandb
from opensora.acceleration.checkpoint import set_grad_checkpoint
from opensora.acceleration.parallel_states import get_data_parallel_group
from opensora.datasets import prepare_dataloader, prepare_variable_dataloader
from opensora.datasets.utils import collate_fn_ignore_none
from opensora.registry import DATASETS, MODELS, SCHEDULERS, build_module
from opensora.utils.ckpt_utils import load, model_gathering, model_sharding, record_model_param_shape, save
from opensora.utils.config_utils import define_experiment_workspace, parse_configs, save_training_config
from opensora.utils.misc import (
    all_reduce_mean,
    create_logger,
    create_tensorboard_writer,
    format_numel_str,
    get_model_numel,
    requires_grad,
    to_torch_dtype,
)
from opensora.utils.train_utils import MaskGenerator, create_colossalai_plugin, update_ema

DEFAULT_DATASET_NAME = "VideoTextDataset"


def main():
    # ======================================================
    # 1. configs & runtime variables
    # ======================================================
    # == parse configs ==
    cfg = parse_configs(training=True)

    # == device and dtype ==
    assert torch.cuda.is_available(), "Training currently requires at least one GPU."
    cfg_dtype = cfg.get("dtype", "bf16")
    assert cfg_dtype in ["fp16", "bf16"], f"Unknown mixed precision {cfg_dtype}"
    dtype = to_torch_dtype(cfg.get("dtype", "bf16"))

    # == colossalai init distributed training ==
    # NOTE: A very large timeout is set to avoid some processes exit early
    dist.init_process_group(backend="nccl", timeout=timedelta(hours=24))
    torch.cuda.set_device(dist.get_rank() % torch.cuda.device_count())
    set_seed(cfg.get("seed", 1024))
    coordinator = DistCoordinator()
    device = get_current_device()

    # == init exp_dir ==
    exp_name, exp_dir = define_experiment_workspace(cfg)
    coordinator.block_all()
    if coordinator.is_master():
        os.makedirs(exp_dir, exist_ok=True)
        save_training_config(cfg.to_dict(), exp_dir)
    coordinator.block_all()

    # == init logger, tensorboard & wandb ==
    logger = create_logger(exp_dir)
    logger.info("Experiment directory created at %s", exp_dir)
    logger.info("Training configuration:\n %s", pformat(cfg.to_dict()))
    if coordinator.is_master():
        tb_writer = create_tensorboard_writer(exp_dir)
        if cfg.get("wandb", False):
            wandb.init(project="minisora", name=exp_name, config=cfg.to_dict(), dir="./outputs/wandb")

    # == init ColossalAI booster ==
    plugin = create_colossalai_plugin(
        plugin=cfg.get("plugin", "zero2"),
        dtype=cfg_dtype,
        grad_clip=cfg.get("grad_clip", 0),
        sp_size=cfg.get("sp_size", 1),
    )
    booster = Booster(plugin=plugin)

    # ======================================================
    # 2. build dataset and dataloader
    # ======================================================
    logger.info("Building dataset...")
    # == build dataset ==
    dataset = build_module(cfg.dataset, DATASETS)
    logger.info("Dataset contains %s samples.", len(dataset))

    # == build dataloader ==
    dataloader_args = dict(
        dataset=dataset,
        batch_size=cfg.get("batch_size", None),
        num_workers=cfg.get("num_workers", 4),
        seed=cfg.get("seed", 1024),
        shuffle=True,
        drop_last=True,
        pin_memory=True,
        process_group=get_data_parallel_group(),
        collate_fn=collate_fn_ignore_none,
    )
    if cfg.dataset.type == DEFAULT_DATASET_NAME:
        dataloader = prepare_dataloader(**dataloader_args)
        total_batch_size = cfg.batch_size * dist.get_world_size() // cfg.get("sp_size", 1)
        logger.info("Total batch size: %s", total_batch_size)
        num_steps_per_epoch = len(dataloader)
        sampler_to_io = None
    else:
        dataloader = prepare_variable_dataloader(
            bucket_config=cfg.get("bucket_config", None),
            num_bucket_build_workers=cfg.get("num_bucket_build_workers", 1),
            **dataloader_args,
        )
        num_steps_per_epoch = dataloader.batch_sampler.get_num_batch() // dist.get_world_size()
        sampler_to_io = None if cfg.get("start_from_scratch ", False) else dataloader.batch_sampler

    # ======================================================
    # 3. build model
    # ======================================================
    logger.info("Building models...")
    # == build text-encoder and vae ==
    text_encoder = build_module(cfg.text_encoder, MODELS, device=device)
    vae = build_module(cfg.vae, MODELS).to(device, dtype).eval()

    # == build diffusion model ==
    input_size = (dataset.num_frames, *dataset.image_size)
    latent_size = vae.get_latent_size(input_size)
    model = (
        build_module(
            cfg.model,
            MODELS,
            input_size=latent_size,
            in_channels=vae.out_channels,
            caption_channels=text_encoder.output_dim,
            model_max_length=text_encoder.model_max_length,
        )
        .to(device, dtype)
        .train()
    )
    model_numel, model_numel_trainable = get_model_numel(model)
    logger.info(
        "[Diffusion] Trainable model params: %s, Total model params: %s",
        format_numel_str(model_numel_trainable),
        format_numel_str(model_numel),
    )

    # == build ema for diffusion model ==
    ema = deepcopy(model).to(torch.float32).to(device)
    requires_grad(ema, False)
    ema_shape_dict = record_model_param_shape(ema)
    ema.eval()
    update_ema(ema, model, decay=0, sharded=False)

    # == setup loss function, build scheduler ==
    scheduler = build_module(cfg.scheduler, SCHEDULERS)

    # == setup optimizer ==
    optimizer = HybridAdam(
        filter(lambda p: p.requires_grad, model.parameters()),
        adamw_mode=True,
        lr=cfg.get("lr", 1e-4),
        weight_decay=cfg.get("weight_decay", 0),
        eps=cfg.get("adam_eps", 1e-8),
    )
    lr_scheduler = None

    # == additional preparation ==
    if cfg.get("grad_checkpoint", False):
        set_grad_checkpoint(model)
    if cfg.get("mask_ratios", None) is not None:
        mask_generator = MaskGenerator(cfg.mask_ratios)

    # =======================================================
    # 4. distributed training preparation with colossalai
    # =======================================================
    logger.info("Preparing for distributed training...")
    # == boosting ==
    # NOTE: we set dtype first to make initialization of model consistent with the dtype; then reset it to the fp32 as we make diffusion scheduler in fp32
    torch.set_default_dtype(dtype)
    model, optimizer, _, dataloader, lr_scheduler = booster.boost(
        model=model,
        optimizer=optimizer,
        lr_scheduler=lr_scheduler,
        dataloader=dataloader,
    )
    torch.set_default_dtype(torch.float)
    logger.info("Boosting model for distributed training")

    # == global variables ==
    cfg_epochs = cfg.get("epochs", 1000)
    start_epoch = start_step = log_step = sampler_start_idx = acc_step = 0
    running_loss = 0.0
    logger.info("Training for %s epochs with %s steps per epoch", cfg_epochs, num_steps_per_epoch)

    # == resume ==
    if cfg.get("load", None) is not None:
        logger.info("Loading checkpoint")
        ret = load(
            booster,
            cfg.load,
            model=model,
            ema=ema,
            optimizer=optimizer,
            lr_scheduler=lr_scheduler,
            sampler=sampler_to_io,
        )
        if not cfg.get("start_from_scratch ", False):
            start_epoch, start_step, sampler_start_idx = ret
        logger.info("Loaded checkpoint %s at epoch %s step %s", cfg.load, start_epoch, start_step)
    if cfg.dataset.type == DEFAULT_DATASET_NAME:
        dataloader.sampler.set_start_index(sampler_start_idx)

    model_sharding(ema)

    # =======================================================
    # 5. training loop
    # =======================================================
    dist.barrier()
    for epoch in range(start_epoch, cfg_epochs):
        # == set dataloader to new epoch ==
        if cfg.dataset.type == DEFAULT_DATASET_NAME:
            dataloader.sampler.set_epoch(epoch)
        dataloader_iter = iter(dataloader)
        logger.info("Beginning epoch %s...", epoch)

        # == training loop in an epoch ==
        with tqdm(
            enumerate(dataloader_iter, start=start_step),
            desc=f"Epoch {epoch}",
            disable=not coordinator.is_master(),
            initial=start_step,
            total=num_steps_per_epoch,
        ) as pbar:
            for step, batch in pbar:
                x = batch.pop("video").to(device, dtype)  # [B, C, T, H, W]
                y = batch.pop("text")

                # == visual and text encoding ==
                with torch.no_grad():
                    # Prepare visual inputs
                    x = vae.encode(x)  # [B, C, T, H/P, W/P]
                    # Prepare text inputs
                    model_args = text_encoder.encode(y)

                # == mask ==
                mask = None
                if cfg.get("mask_ratios", None) is not None:
                    mask = mask_generator.get_masks(x)
                    model_args["x_mask"] = mask

                # == video meta info ==
                for k, v in batch.items():
                    model_args[k] = v.to(device, dtype)

                # == diffusion loss computation ==
                loss_dict = scheduler.training_losses(model, x, model_args, mask=mask)

                # == backward & update ==
                loss = loss_dict["loss"].mean()
                booster.backward(loss=loss, optimizer=optimizer)
                optimizer.step()
                optimizer.zero_grad()

                # == update EMA ==
                update_ema(ema, model.module, optimizer=optimizer, decay=cfg.get("ema_decay", 0.9999))

                # == update log info ==
                all_reduce_mean(loss)
                running_loss += loss.item()
                global_step = epoch * num_steps_per_epoch + step
                log_step += 1
                acc_step += 1

                # == logging ==
                if coordinator.is_master() and (global_step + 1) % cfg.get("log_every", 1) == 0:
                    avg_loss = running_loss / log_step
                    # progress bar
                    pbar.set_postfix({"loss": avg_loss, "step": step, "global_step": global_step})
                    # tensorboard
                    tb_writer.add_scalar("loss", loss.item(), global_step)
                    # wandb
                    if cfg.get("wandb", False):
                        wandb.log(
                            {
                                "iter": global_step,
                                "epoch": epoch,
                                "loss": loss.item(),
                                "avg_loss": avg_loss,
                                "acc_step": acc_step,
                            },
                            step=global_step,
                        )

                    running_loss = 0.0
                    log_step = 0

                # == checkpoint saving ==
                ckpt_every = cfg.get("ckpt_every", 0)
                if ckpt_every > 0 and (global_step + 1) % ckpt_every == 0:
                    model_gathering(ema, ema_shape_dict)
                    save(
                        booster,
                        exp_dir,
                        model=model,
                        ema=ema,
                        optimizer=optimizer,
                        lr_scheduler=lr_scheduler,
                        sampler=sampler_to_io,
                        epoch=epoch,
                        step=step + 1,
                        global_step=global_step + 1,
                        batch_size=cfg.get("batch_size", None),
                    )
                    if dist.get_rank() == 0:
                        model_sharding(ema)
                    logger.info(
                        "Saved checkpoint at epoch %s step %s global_step %s to %s",
                        epoch,
                        step + 1,
                        global_step + 1,
                        exp_dir,
                    )

        # NOTE: the continue epochs are not resumed, so we need to reset the sampler start index and start step
        if cfg.dataset.type == DEFAULT_DATASET_NAME:
            dataloader.sampler.set_start_index(0)
        else:
            dataloader.batch_sampler.set_epoch(epoch + 1)
            logger.info("Epoch done, recomputing batch sampler")
        start_step = 0


if __name__ == "__main__":
    main()