train_ppt2_bn.py

#!/usr/bin/env python
# modified from https://github.com/TencentARC/BrushNet/blob/main/examples/brushnet/train_brushnet.py

import argparse
import gc
import logging
import math
import os
import shutil
from pathlib import Path

import accelerate
import numpy as np
import torch
import torch.nn.functional as F
import torch.utils.checkpoint
import transformers
from accelerate import Accelerator
from accelerate.logging import get_logger
from accelerate.utils import ProjectConfiguration, set_seed
from huggingface_hub import create_repo, upload_folder
from omegaconf import OmegaConf
from packaging import version
from PIL import Image
from torchvision import transforms
from tqdm.auto import tqdm
from transformers import PretrainedConfig

import diffusers
import powerpaint.datasets
from diffusers.optimization import get_scheduler
from diffusers.training_utils import compute_snr
from diffusers.utils import check_min_version, is_wandb_available
from diffusers.utils.hub_utils import load_or_create_model_card, populate_model_card
from diffusers.utils.import_utils import is_xformers_available
from diffusers.utils.torch_utils import is_compiled_module
from powerpaint.datasets import ProbPickingDataset
from powerpaint.models import BrushNetModel, UNet2DConditionModel
from powerpaint.pipelines import StableDiffusionPowerPaintBrushNetPipeline


if is_wandb_available():
    import wandb

# Will error if the minimal version of diffusers is not installed. Remove at your own risks.
check_min_version("0.27.0.dev0")

logger = get_logger(__name__)


def save_model_card(repo_id: str, image_logs=None, base_model=str, repo_folder=None):
    img_str = ""
    if image_logs is not None:
        img_str = "You can find some example images below.\n\n"
        for i, log in enumerate(image_logs):
            images = log["images"]
            validation_prompt = log["validation_prompt"]
            validation_image = log["validation_image"]

            validation_image.save(os.path.join(repo_folder, f"image_{i}.png"))
            img_str += f"prompt: {validation_prompt}\n"
            images = [validation_image] + images
            img_str += f"![images_{i})](./images_{i}.png)\n"

    model_description = f"""
# PowerPaint - {repo_id}

These are PowerPaint weights trained on {base_model} with new type of conditioning.
{img_str}
"""
    model_card = load_or_create_model_card(
        repo_id_or_path=repo_id,
        from_training=True,
        license="creativeml-openrail-m",
        base_model=base_model,
        model_description=model_description,
        inference=True,
    )

    tags = [
        "stable-diffusion",
        "stable-diffusion-diffusers",
        "text-to-image",
        "diffusers",
        "PowerPaint",
        "diffusers-training",
    ]
    model_card = populate_model_card(model_card, tags=tags)

    model_card.save(os.path.join(repo_folder, "README.md"))


def log_validation(tokenizer, text_encoder, brushnet, args, accelerator, weight_dtype, step):
    logger.info("Running validation... ")

    # use fixed model from pretrained models, and text_encoder and tokenizer from trainer
    pipe = StableDiffusionPowerPaintBrushNetPipeline.from_pretrained(
        args.pretrained_model_name_or_path,
        unet=UNet2DConditionModel.from_pretrained(
            args.pretrained_model_name_or_path, subfolder="unet", revision=args.revision, variant=args.variant
        ),
        tokenizer=tokenizer,
        text_encoder=accelerator.unwrap_model(text_encoder),
        brushnet=accelerator.unwrap_model(brushnet),
        safety_checker=None,
        revision=args.revision,
        variant=args.variant,
        torch_dtype=weight_dtype,
        local_files_only=True,  # load files from local cache
    )

    pipe = pipe.to(accelerator.device)
    pipe.set_progress_bar_config(disable=True)

    if args.enable_xformers_memory_efficient_attention:
        pipe.enable_xformers_memory_efficient_attention()

    # load validation images
    image_logs = []
    for case in args.validation_data.cases:
        validation_prompts = case.prompt
        validation_image = Image.open(os.path.join(args.validation_data.data_root, case.image)).convert("RGB")
        validation_mask = Image.open(os.path.join(args.validation_data.data_root, case.mask))
        validation_mask = validation_mask.resize((validation_image.size[0], validation_image.size[1]), Image.NEAREST)
        validation_mask = validation_mask.convert("L")
        hole_value = (0, 0, 0)
        validation_image = Image.composite(
            Image.new("RGB", (validation_image.size[0], validation_image.size[1]), hole_value),
            validation_image,
            validation_mask.convert("L"),
        )

        image_grid = Image.new(
            "RGB",
            (validation_image.size[0] * (1 + len(validation_prompts)), validation_image.size[1]),
            (255, 255, 255),
        )
        image_grid.paste(validation_image, (0, 0))
        t2i_mask = Image.new("RGB", (validation_image.size[0], validation_image.size[1]), (255, 255, 255)).convert("L")
        t2i_image = Image.new("RGB", (validation_image.size[0], validation_image.size[1]), (0, 0, 0))
        for i, p in enumerate(validation_prompts):
            with torch.autocast(accelerator.device.type):
                image = pipe(
                    promptA=p.promptA,
                    promptB=p.promptB,
                    prompt=p.prompt,
                    negative_promptA=p.negative_promptA,
                    negative_promptB=p.negative_promptB,
                    negative_prompt=p.negative_prompt,
                    tradeoff=p.tradeoff,
                    image=validation_image if p.task != "t2i" else t2i_image,
                    mask=validation_mask if p.task != "t2i" else t2i_mask,
                    num_inference_steps=20,
                ).images[0]
            image_logs.append(image)
            image_grid.paste(image, (validation_image.size[0] * (i + 1), 0))
        image_grid.save(os.path.join(args.output_dir, f"{str(step).zfill(3)}_{os.path.basename(case.image)}"))
    gc.collect()
    torch.cuda.empty_cache()

    for tracker in accelerator.trackers:
        if tracker.name == "tensorboard":
            np_images = np.stack([np.asarray(img) for img in image_logs])
            tracker.writer.add_images("validation", np_images, step, dataformats="NHWC")
        elif tracker.name == "wandb":
            tracker.log(
                {
                    "validation": [
                        wandb.Image(image, caption=f"{p.task}")
                        for image, p in zip(image_logs, args.validation_data.cases[0].prompt)
                    ]
                }
            )
        else:
            logger.warning(f"image logging not implemented for {tracker.name}")

    del pipe
    gc.collect()
    torch.cuda.empty_cache()

    return image_logs


def import_model_class_from_model_name_or_path(pretrained_model_name_or_path: str, revision: str):
    text_encoder_config = PretrainedConfig.from_pretrained(
        pretrained_model_name_or_path,
        subfolder="text_encoder",
        revision=revision,
    )
    model_class = text_encoder_config.architectures[0]

    if model_class == "CLIPTextModel":
        from transformers import CLIPTextModel

        return CLIPTextModel
    elif model_class == "RobertaSeriesModelWithTransformation":
        from diffusers.pipelines.alt_diffusion.modeling_roberta_series import RobertaSeriesModelWithTransformation

        return RobertaSeriesModelWithTransformation
    elif model_class == "T5EncoderModel":
        from transformers import T5EncoderModel

        return T5EncoderModel
    else:
        raise ValueError(f"{model_class} is not supported.")


def parse_args(input_args=None):
    parser = argparse.ArgumentParser(
        description="Simple example of a PowerPaint based on brushnet architecture training script."
    )

    parser.add_argument(
        "--config",
        type=str,
        default=None,
        help="yaml for configuration",
    )
    parser.add_argument(
        "--pretrained_model_name_or_path",
        type=str,
        default=None,
        required=False,
        help="Path to pretrained model or model identifier from huggingface.co/models.",
    )
    parser.add_argument(
        "--powerpaint_model_name_or_path",
        type=str,
        default=None,
        help="Path to pretrained powerpaint model or model identifier from huggingface.co/models."
        " If not specified powerpaint weights are initialized from unet.",
    )
    parser.add_argument(
        "--output_dir",
        type=str,
        default="runs/ppt2_bn",
        help="The output directory where the model predictions and checkpoints will be written.",
    )
    parser.add_argument(
        "--revision",
        type=str,
        default=None,
        required=False,
        help="Revision of pretrained model identifier from huggingface.co/models.",
    )
    parser.add_argument(
        "--variant",
        type=str,
        default=None,
        help="Variant of the model files of the pretrained model identifier from huggingface.co/models, 'e.g.' fp16",
    )
    parser.add_argument(
        "--tokenizer_name",
        type=str,
        default=None,
        help="Pretrained tokenizer name or path if not the same as model_name",
    )
    parser.add_argument(
        "--cache_dir",
        type=str,
        default=None,
        help="The directory where the downloaded models and datasets will be stored.",
    )
    parser.add_argument("--seed", type=int, default=None, help="A seed for reproducible training.")
    parser.add_argument(
        "--resolution",
        type=int,
        default=512,
        help=(
            "The resolution for input images, all the images in the train/validation dataset will be resized to this"
            " resolution"
        ),
    )
    parser.add_argument(
        "--train_batch_size", type=int, default=4, help="Batch size (per device) for the training dataloader."
    )
    parser.add_argument("--num_train_epochs", type=int, default=10000)
    parser.add_argument(
        "--max_train_steps",
        type=int,
        default=None,
        help="Total number of training steps to perform.  If provided, overrides num_train_epochs.",
    )
    parser.add_argument(
        "--checkpointing_steps",
        type=int,
        default=500,
        help=(
            "Save a checkpoint of the training state every X updates. Checkpoints can be used for resuming training via `--resume_from_checkpoint`. "
            "In the case that the checkpoint is better than the final trained model, the checkpoint can also be used for inference."
            "Using a checkpoint for inference requires separate loading of the original pipeline and the individual checkpointed model components."
            "See https://huggingface.co/docs/diffusers/main/en/training/dreambooth#performing-inference-using-a-saved-checkpoint for step by step"
            "instructions."
        ),
    )
    parser.add_argument(
        "--checkpoints_total_limit",
        type=int,
        default=None,
        help=("Max number of checkpoints to store."),
    )
    parser.add_argument(
        "--resume_from_checkpoint",
        type=str,
        default=None,
        help=(
            "Whether training should be resumed from a previous checkpoint. Use a path saved by"
            ' `--checkpointing_steps`, or `"latest"` to automatically select the last available checkpoint.'
        ),
    )
    parser.add_argument(
        "--gradient_accumulation_steps",
        type=int,
        default=1,
        help="Number of updates steps to accumulate before performing a backward/update pass.",
    )
    parser.add_argument(
        "--gradient_checkpointing",
        action="store_true",
        help="Whether or not to use gradient checkpointing to save memory at the expense of slower backward pass.",
    )
    parser.add_argument(
        "--learning_rate",
        type=float,
        default=5e-6,
        help="Initial learning rate (after the potential warm up period) to use.",
    )
    parser.add_argument(
        "--scale_lr",
        action="store_true",
        default=False,
        help="Scale the learning rate by the number of GPUs, gradient accumulation steps, and batch size.",
    )
    parser.add_argument(
        "--lr_scheduler",
        type=str,
        default="constant",
        help=(
            'The scheduler type to use. Choose between ["linear", "cosine", "cosine_with_restarts", "polynomial",'
            ' "constant", "constant_with_warmup"]'
        ),
    )
    parser.add_argument(
        "--lr_warmup_steps", type=int, default=500, help="Number of steps for the warm up in the lr scheduler."
    )
    parser.add_argument(
        "--lr_num_cycles",
        type=int,
        default=1,
        help="Number of hard resets of the lr in cosine_with_restarts scheduler.",
    )
    parser.add_argument("--lr_power", type=float, default=1.0, help="Power factor of the polynomial scheduler.")
    parser.add_argument(
        "--use_8bit_adam", action="store_true", help="Whether or not to use 8-bit Adam from bitsandbytes."
    )
    parser.add_argument(
        "--dataloader_num_workers",
        type=int,
        default=0,
        help=(
            "Number of subprocesses to use for data loading. 0 means that the data will be loaded in the main process."
        ),
    )
    parser.add_argument("--adam_beta1", type=float, default=0.9, help="The beta1 parameter for the Adam optimizer.")
    parser.add_argument("--adam_beta2", type=float, default=0.999, help="The beta2 parameter for the Adam optimizer.")
    parser.add_argument("--adam_weight_decay", type=float, default=1e-2, help="Weight decay to use.")
    parser.add_argument("--adam_epsilon", type=float, default=1e-08, help="Epsilon value for the Adam optimizer")
    parser.add_argument("--max_grad_norm", default=1.0, type=float, help="Max gradient norm.")
    parser.add_argument("--push_to_hub", action="store_true", help="Whether or not to push the model to the Hub.")
    parser.add_argument("--hub_token", type=str, default=None, help="The token to use to push to the Model Hub.")
    parser.add_argument(
        "--hub_model_id",
        type=str,
        default=None,
        help="The name of the repository to keep in sync with the local `output_dir`.",
    )
    parser.add_argument(
        "--logging_dir",
        type=str,
        default="logs",
        help=(
            "[TensorBoard](https://www.tensorflow.org/tensorboard) log directory. Will default to"
            " *output_dir/runs/**CURRENT_DATETIME_HOSTNAME***."
        ),
    )
    parser.add_argument(
        "--allow_tf32",
        action="store_true",
        help=(
            "Whether or not to allow TF32 on Ampere GPUs. Can be used to speed up training. For more information, see"
            " https://pytorch.org/docs/stable/notes/cuda.html#tensorfloat-32-tf32-on-ampere-devices"
        ),
    )
    parser.add_argument(
        "--report_to",
        type=str,
        default="tensorboard",
        help=(
            'The integration to report the results and logs to. Supported platforms are `"tensorboard"`'
            ' (default), `"wandb"` and `"comet_ml"`. Use `"all"` to report to all integrations.'
        ),
    )
    parser.add_argument(
        "--mixed_precision",
        type=str,
        default=None,
        choices=["no", "fp16", "bf16"],
        help=(
            "Whether to use mixed precision. Choose between fp16 and bf16 (bfloat16). Bf16 requires PyTorch >="
            " 1.10.and an Nvidia Ampere GPU.  Default to the value of accelerate config of the current system or the"
            " flag passed with the `accelerate.launch` command. Use this argument to override the accelerate config."
        ),
    )
    parser.add_argument(
        "--enable_xformers_memory_efficient_attention", action="store_true", help="Whether or not to use xformers."
    )
    parser.add_argument(
        "--set_grads_to_none",
        action="store_true",
        help=(
            "Save more memory by using setting grads to None instead of zero. Be aware, that this changes certain"
            " behaviors, so disable this argument if it causes any problems. More info:"
            " https://pytorch.org/docs/stable/generated/torch.optim.Optimizer.zero_grad.html"
        ),
    )
    parser.add_argument(
        "--dataset_name",
        type=str,
        default=None,
        help=(
            "The name of the Dataset (from the HuggingFace hub) to train on (could be your own, possibly private,"
            " dataset). It can also be a path pointing to a local copy of a dataset in your filesystem,"
            " or to a folder containing files that 🤗 Datasets can understand."
        ),
    )
    parser.add_argument(
        "--dataset_config_name",
        type=str,
        default=None,
        help="The config of the Dataset, leave as None if there's only one config.",
    )

    parser.add_argument(
        "--image_column", type=str, default="image", help="The column of the dataset containing the target image."
    )
    parser.add_argument(
        "--conditioning_image_column",
        type=str,
        default="conditioning_image",
        help="The column of the dataset containing the powerpaint conditioning image.",
    )
    parser.add_argument(
        "--caption_column",
        type=str,
        default="text",
        help="The column of the dataset containing a caption or a list of captions.",
    )
    parser.add_argument(
        "--max_train_samples",
        type=int,
        default=None,
        help=(
            "For debugging purposes or quicker training, truncate the number of training examples to this "
            "value if set."
        ),
    )
    parser.add_argument(
        "--proportion_empty_prompts",
        type=float,
        default=0,
        help="Proportion of image prompts to be replaced with empty strings. Defaults to 0 (no prompt replacement).",
    )
    parser.add_argument(
        "--snr_gamma",
        type=float,
        default=None,
        help="SNR weighting gamma to be used if rebalancing the loss. Recommended value is 5.0. "
        "More details here: https://arxiv.org/abs/2303.09556.",
    )
    parser.add_argument(
        "--validation_steps",
        type=int,
        default=100,
        help=(
            "Run validation every X steps. Validation consists of running the prompt"
            " `args.validation_prompt` multiple times: `args.num_validation_images`"
            " and logging the images."
        ),
    )
    parser.add_argument(
        "--tracker_project_name",
        type=str,
        default="train_powerpaint_brushnet",
        help=(
            "The `project_name` argument passed to Accelerator.init_trackers for"
            " more information see https://huggingface.co/docs/accelerate/v0.17.0/en/package_reference/accelerator#accelerate.Accelerator"
        ),
    )

    if input_args is not None:
        args = parser.parse_args(input_args)
    else:
        args = parser.parse_args()

    # use omegaconf to manage configurations
    if args.config is not None:
        config = OmegaConf.load(args.config)
        for k, v in config.items():
            args.__dict__[k] = v

    if args.proportion_empty_prompts < 0 or args.proportion_empty_prompts > 1:
        raise ValueError("`--proportion_empty_prompts` must be in the range [0, 1].")

    if args.resolution % 8 != 0:
        raise ValueError(
            "`--resolution` must be divisible by 8 for consistently sized encoded images between the VAE and the brushnet encoder."
        )

    return args


def main(args):
    if args.report_to == "wandb" and args.hub_token is not None:
        raise ValueError(
            "You cannot use both --report_to=wandb and --hub_token due to a security risk of exposing your token."
            " Please use `huggingface-cli login` to authenticate with the Hub."
        )

    logging_dir = Path(args.output_dir, args.logging_dir)
    accelerator_project_config = ProjectConfiguration(project_dir=args.output_dir, logging_dir=logging_dir)
    accelerator = Accelerator(
        gradient_accumulation_steps=args.gradient_accumulation_steps,
        mixed_precision=args.mixed_precision,
        log_with=args.report_to,
        project_config=accelerator_project_config,
    )

    # Make one log on every process with the configuration for debugging.
    logging.basicConfig(
        format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
        datefmt="%m/%d/%Y %H:%M:%S",
        level=logging.INFO,
    )
    logger.info(accelerator.state, main_process_only=False)
    if accelerator.is_local_main_process:
        transformers.utils.logging.set_verbosity_warning()
        diffusers.utils.logging.set_verbosity_info()
    else:
        transformers.utils.logging.set_verbosity_error()
        diffusers.utils.logging.set_verbosity_error()

    # If passed along, set the training seed now.
    if args.seed is not None:
        torch.manual_seed(args.seed)
        set_seed(args.seed)

    # Handle the repository creation
    if accelerator.is_main_process:
        if args.output_dir is not None:
            os.makedirs(args.output_dir, exist_ok=True)

        # saving training configuration to output_dir
        to_save_config = OmegaConf.create(vars(args))
        OmegaConf.save(config=to_save_config, f=os.path.join(args.output_dir, "training_config.yaml"))

        if args.push_to_hub:
            repo_id = create_repo(
                repo_id=args.hub_model_id or Path(args.output_dir).name, exist_ok=True, token=args.hub_token
            ).repo_id

    # For mixed precision training we cast the text_encoder and vae weights to half-precision
    # as these models are only used for inference, keeping weights in full precision is not required.
    weight_dtype = torch.float32
    if accelerator.mixed_precision == "fp16":
        weight_dtype = torch.float16
    elif accelerator.mixed_precision == "bf16":
        weight_dtype = torch.bfloat16

    # initialize from pre-trained pipeline
    pipe = StableDiffusionPowerPaintBrushNetPipeline.from_pretrained(
        args.pretrained_model_name_or_path,
        unet=UNet2DConditionModel.from_pretrained(
            args.pretrained_model_name_or_path, subfolder="unet", revision=args.revision, variant=args.variant
        ),
        safety_checker=None,
        revision=args.revision,
        variant=args.variant,
        torch_dtype=weight_dtype,
        local_files_only=True,  # load files from local cache
    )

    if args.powerpaint_model_name_or_path:
        logger.info("Loading existing powerpaint weights")
        pipe.brushnet = BrushNetModel.from_pretrained(args.powerpaint_model_name_or_path)

    # Taken from [Sayak Paul's Diffusers PR #6511](https://github.com/huggingface/diffusers/pull/6511/files)
    def unwrap_model(model):
        model = accelerator.unwrap_model(model)
        model = model._orig_mod if is_compiled_module(model) else model
        return model

    # IMPORTANT: add learnable tokens for task prompts into tokenizer
    placeholder_tokens = [v.placeholder_tokens for k, v in args.task_prompt.items()]
    initializer_token = [v.initializer_token for k, v in args.task_prompt.items()]
    num_vectors_per_token = [v.num_vectors_per_token for k, v in args.task_prompt.items()]
    placeholder_token_ids = pipe.add_tokens(
        placeholder_tokens, initializer_token, num_vectors_per_token, initialize_parameters=True
    )

    vae, tokenizer, unet, noise_scheduler = pipe.vae, pipe.tokenizer, pipe.unet, pipe.scheduler
    text_encoder, brushnet = pipe.text_encoder.to(torch.float32), pipe.brushnet.to(torch.float32)
    text_encoder_cls = import_model_class_from_model_name_or_path(args.pretrained_model_name_or_path, args.revision)

    # `accelerate` 0.16.0 will have better support for customized saving
    if version.parse(accelerate.__version__) >= version.parse("0.16.0"):
        # create custom saving & loading hooks so that `accelerator.save_state(...)` serializes in a nice format
        def save_model_hook(models, weights, output_dir):
            if accelerator.is_main_process:
                for model in models:
                    sub_dir = "brushnet" if isinstance(model, type(unwrap_model(brushnet))) else "text_encoder"
                    model.save_pretrained(os.path.join(output_dir, sub_dir))

                    # make sure to pop weight so that corresponding model is not saved again
                    weights.pop()

        def load_model_hook(models, input_dir):
            while len(models) > 0:
                model = models.pop()

                if isinstance(model, type(unwrap_model(text_encoder))):
                    # load transformers style into model
                    load_model = text_encoder_cls.from_pretrained(input_dir, subfolder="text_encoder")
                    model.config = load_model.config
                else:
                    # load diffusers style into model
                    load_model = BrushNetModel.from_pretrained(input_dir, subfolder="brushnet")
                    model.register_to_config(**load_model.config)

                model.load_state_dict(load_model.state_dict())
                del load_model

        accelerator.register_save_state_pre_hook(save_model_hook)
        accelerator.register_load_state_pre_hook(load_model_hook)

    if args.gradient_checkpointing:
        brushnet.enable_gradient_checkpointing()

    if args.enable_xformers_memory_efficient_attention:
        if is_xformers_available():
            import xformers

            xformers_version = version.parse(xformers.__version__)
            if xformers_version == version.parse("0.0.16"):
                logger.warn(
                    "xFormers 0.0.16 cannot be used for training in some GPUs. If you observe problems during training, please update xFormers to at least 0.0.17. See https://huggingface.co/docs/diffusers/main/en/optimization/xformers for more details."
                )
            unet.enable_xformers_memory_efficient_attention()
            brushnet.enable_xformers_memory_efficient_attention()
        else:
            raise ValueError("xformers is not available. Make sure it is installed correctly")

    # Check that all trainable models are in full precision
    low_precision_error_string = (
        " Please make sure to always have all model weights in full float32 precision when starting training - even if"
        " doing mixed precision training, copy of the weights should still be float32."
    )

    if unwrap_model(brushnet).dtype != torch.float32:
        raise ValueError(f"BrushNet loaded as datatype {unwrap_model(brushnet).dtype}. {low_precision_error_string}")

    # Enable TF32 for faster training on Ampere GPUs,
    # cf https://pytorch.org/docs/stable/notes/cuda.html#tensorfloat-32-tf32-on-ampere-devices
    if args.allow_tf32:
        torch.backends.cuda.matmul.allow_tf32 = True

    if args.scale_lr:
        args.learning_rate = (
            args.learning_rate * args.gradient_accumulation_steps * args.train_batch_size * accelerator.num_processes
        )

    # Use 8-bit Adam for lower memory usage or to fine-tune the model in 16GB GPUs
    if args.use_8bit_adam:
        try:
            import bitsandbytes as bnb
        except ImportError:
            raise ImportError(
                "To use 8-bit Adam, please install the bitsandbytes library: `pip install bitsandbytes`."
            )

        optimizer_class = bnb.optim.AdamW8bit
    else:
        optimizer_class = torch.optim.AdamW

    # 1. trainable embedding + 2. trainable model (brushnet)
    vae.requires_grad_(False)
    unet.requires_grad_(False)

    # Freeze all parameters except for the token embeddings in text encoder
    text_encoder.text_model.encoder.requires_grad_(False)
    text_encoder.text_model.final_layer_norm.requires_grad_(False)
    text_encoder.text_model.embeddings.position_embedding.requires_grad_(False)

    optimizer = optimizer_class(
        list(brushnet.parameters()) + list(text_encoder.get_input_embeddings().parameters()),
        lr=args.learning_rate,
        betas=(args.adam_beta1, args.adam_beta2),
        weight_decay=args.adam_weight_decay,
        eps=args.adam_epsilon,
    )

    # transforms used for preprocessing dataset
    train_transforms = transforms.Compose(
        [
            transforms.Resize(args.resolution, interpolation=transforms.InterpolationMode.BILINEAR),
            transforms.CenterCrop(args.resolution),
            transforms.RandomHorizontalFlip(),
            transforms.ToTensor(),
            transforms.Normalize([0.5], [0.5]),
        ]
    )

    # preparing datasets and dataloader for training.
    # support loading multiple datasets in a single dataloader.
    datasets_list = []
    for d in args.train_data.datasets:
        dataset_class = getattr(powerpaint.datasets, d.dataset_class)
        dataset_ = dataset_class(train_transforms, pipe, args.task_prompt, **d)
        datasets_list.append({"dataset": dataset_, "prob": d.prob})

    train_dataset = ProbPickingDataset(datasets_list)

    with accelerator.main_process_first():
        if args.max_train_samples is not None:
            train_dataset = train_dataset.shuffle(seed=args.seed).select(range(args.max_train_samples))

    train_dataloader = torch.utils.data.DataLoader(
        train_dataset,
        batch_size=args.train_batch_size,
        num_workers=args.dataloader_num_workers,
    )

    # Scheduler and math around the number of training steps.
    overrode_max_train_steps = False
    num_update_steps_per_epoch = math.ceil(len(train_dataloader) / args.gradient_accumulation_steps)
    if args.max_train_steps is None:
        args.max_train_steps = args.num_train_epochs * num_update_steps_per_epoch
        overrode_max_train_steps = True

    lr_scheduler = get_scheduler(
        args.lr_scheduler,
        optimizer=optimizer,
        num_warmup_steps=args.lr_warmup_steps * accelerator.num_processes,
        num_training_steps=args.max_train_steps * accelerator.num_processes,
        num_cycles=args.lr_num_cycles,
        power=args.lr_power,
    )

    brushnet.train()
    text_encoder.train()
    # Prepare everything with our `accelerator`.
    brushnet, text_encoder, optimizer, train_dataloader, lr_scheduler = accelerator.prepare(
        brushnet, text_encoder, optimizer, train_dataloader, lr_scheduler
    )

    # Move vae, unet and text_encoder to device and cast to weight_dtype
    vae.to(accelerator.device, dtype=weight_dtype)
    unet.to(accelerator.device, dtype=weight_dtype)

    # We need to recalculate our total training steps as the size of the training dataloader may have changed.
    num_update_steps_per_epoch = math.ceil(len(train_dataloader) / args.gradient_accumulation_steps)
    if overrode_max_train_steps:
        args.max_train_steps = args.num_train_epochs * num_update_steps_per_epoch
    # Afterwards we recalculate our number of training epochs
    args.num_train_epochs = math.ceil(args.max_train_steps / num_update_steps_per_epoch)

    # We need to initialize the trackers we use, and also store our configuration.
    # The trackers initializes automatically on the main process.
    if accelerator.is_main_process:
        tracker_config = dict(vars(args))

        # tensorboard cannot handle list types for config
        pop_list = []
        for k, v in tracker_config.items():
            if not isinstance(v, (int, float, str, bool, torch.Tensor)):
                pop_list.append(k)
                logger.info(f"Removed {k} (type:{type(v)}) from tracker_config")
        for k in pop_list:
            tracker_config.pop(k)

        accelerator.init_trackers(args.tracker_project_name, config=tracker_config)

    # Train!
    total_batch_size = args.train_batch_size * accelerator.num_processes * args.gradient_accumulation_steps

    logger.info(f"***** Running training for {args.tracker_project_name} *****")
    logger.info(f"  Num examples = {len(train_dataset)}")
    logger.info(f"  Num Epochs = {args.num_train_epochs}")
    logger.info(f"  Instantaneous batch size per device = {args.train_batch_size}")
    logger.info(f"  Total train batch size (w. parallel, distributed & accumulation) = {total_batch_size}")
    logger.info(f"  Gradient Accumulation steps = {args.gradient_accumulation_steps}")
    logger.info(f"  Total optimization steps = {int(args.max_train_steps)}")
    global_step = 0
    first_epoch = 0

    # Potentially load in the weights and states from a previous save
    if args.resume_from_checkpoint:
        if args.resume_from_checkpoint != "latest":
            path = os.path.basename(args.resume_from_checkpoint)
        else:
            # Get the most recent checkpoint
            dirs = os.listdir(args.output_dir)
            dirs = [d for d in dirs if d.startswith("checkpoint")]
            dirs = sorted(dirs, key=lambda x: int(x.split("-")[1]))
            path = dirs[-1] if len(dirs) > 0 else None

        if path is None:
            logger.info(f"Checkpoint '{args.resume_from_checkpoint}' does not exist. Starting a new training run.")
            args.resume_from_checkpoint = None
            initial_global_step = 0
        else:
            logger.info(f"Resuming from checkpoint {path}")
            accelerator.load_state(os.path.join(args.output_dir, path), map_location="cpu")
            global_step = int(path.split("-")[1])

            initial_global_step = global_step
            first_epoch = global_step // num_update_steps_per_epoch
    else:
        initial_global_step = 0

    progress_bar = tqdm(
        range(0, int(args.max_train_steps)),
        initial=initial_global_step,
        desc="Steps",
        # Only show the progress bar once on each machine.
        disable=not accelerator.is_local_main_process,
    )

    image_logs = None

    # keep original embeddings as reference
    orig_embeds_params = accelerator.unwrap_model(text_encoder).get_input_embeddings().weight.data.clone()

    for _ in range(first_epoch, args.num_train_epochs):
        train_loss = 0.0
        for batch in train_dataloader:
            with accelerator.accumulate(brushnet):
                # Convert images to latent space
                latents = vae.encode(batch["pixel_values"].to(dtype=weight_dtype)).latent_dist.sample().detach()
                latents = latents * vae.config.scaling_factor

                # we follow the same annotation for mask as
                # https://github.com/huggingface/diffusers/blob/v0.30.0/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion_inpaint.py
                # mask: 1 for masked regions and 0 for known regions
                mask = torch.nn.functional.interpolate(batch["mask"], size=(64, 64))
                mask_image = batch["pixel_values"] * (batch["mask"] < 0.5)
                # convert the hole value from 0 to -1 due to [-1, 1] range
                mask_image = mask_image - batch["mask"]
                mask_image_latents = vae.encode(mask_image.to(weight_dtype)).latent_dist.sample()
                mask_image_latents = (mask_image_latents * vae.config.scaling_factor).to(weight_dtype)

                conditioning_latents = torch.concat([mask, mask_image_latents], 1)

                # Sample noise that we'll add to the latents
                noise = torch.randn_like(latents)
                bsz = latents.shape[0]

                # Sample a random timestep for each image
                timesteps = torch.randint(0, noise_scheduler.config.num_train_timesteps, (bsz,), device=latents.device)
                timesteps = timesteps.long()

                # Add noise to the latents according to the noise magnitude at each timestep
                # (this is the forward diffusion process)
                noisy_latents = noise_scheduler.add_noise(latents, noise, timesteps)

                # Get the text embedding for conditioning unet
                encoder_hidden_states_unet = text_encoder(batch["input_ids"], return_dict=False)[0]

                # text embedding for brushnet, (bs, 77, 768)
                encoder_hidden_statesA = text_encoder(batch["input_idsA"], return_dict=False)[0]
                encoder_hidden_statesB = text_encoder(batch["input_idsB"], return_dict=False)[0]

                # tradeoff between two text embeddings (bs, 2, 1)
                tradeoff = batch["tradeoff"].unsqueeze(-1)
                encoder_hidden_states_brushnet = (
                    tradeoff[:, 0:1, :] * encoder_hidden_statesA + tradeoff[:, 1:, :] * encoder_hidden_statesB.detach()
                )

                # Run the brushnet forward pass
                down_block_res_samples, mid_block_res_sample, up_block_res_samples = brushnet(
                    noisy_latents,
                    timesteps,
                    encoder_hidden_states=encoder_hidden_states_brushnet.to(weight_dtype),
                    brushnet_cond=conditioning_latents,
                    return_dict=False,
                )

                # Predict the noise residual
                model_pred = unet(
                    noisy_latents,
                    timesteps,
                    encoder_hidden_states=encoder_hidden_states_unet.detach().to(weight_dtype),
                    down_block_add_samples=[sample.to(dtype=weight_dtype) for sample in down_block_res_samples],
                    mid_block_add_sample=mid_block_res_sample.to(dtype=weight_dtype),
                    up_block_add_samples=[sample.to(dtype=weight_dtype) for sample in up_block_res_samples],
                    return_dict=False,
                )[0]

                # Get the target for loss depending on the prediction type
                if noise_scheduler.config.prediction_type == "epsilon":
                    target = noise
                elif noise_scheduler.config.prediction_type == "v_prediction":
                    target = noise_scheduler.get_velocity(latents, noise, timesteps)
                else:
                    raise ValueError(f"Unknown prediction type {noise_scheduler.config.prediction_type}")

                if args.snr_gamma is None:
                    loss = F.mse_loss(model_pred.float(), target.float(), reduction="mean")
                else:
                    # Compute loss-weights as per Section 3.4 of https://arxiv.org/abs/2303.09556.
                    # Since we predict the noise instead of x_0, the original formulation is slightly changed.
                    # This is discussed in Section 4.2 of the same paper.
                    snr = compute_snr(noise_scheduler, timesteps)
                    mse_loss_weights = torch.stack([snr, args.snr_gamma * torch.ones_like(timesteps)], dim=1).min(
                        dim=1
                    )[0]
                    if noise_scheduler.config.prediction_type == "epsilon":
                        mse_loss_weights = mse_loss_weights / snr
                    elif noise_scheduler.config.prediction_type == "v_prediction":
                        mse_loss_weights = mse_loss_weights / (snr + 1)

                    loss = F.mse_loss(model_pred.float(), target.float(), reduction="none")
                    loss = loss.mean(dim=list(range(1, len(loss.shape)))) * mse_loss_weights
                    loss = loss.mean()

                # Gather the losses across all processes for logging (if we use distributed training).
                avg_loss = accelerator.gather(loss.repeat(args.train_batch_size)).mean()
                train_loss += avg_loss.item() / args.gradient_accumulation_steps

                accelerator.backward(loss)
                if accelerator.sync_gradients:
                    params_to_clip = list(brushnet.parameters()) + list(
                        accelerator.unwrap_model(text_encoder).get_input_embeddings().parameters()
                    )
                    accelerator.clip_grad_norm_(params_to_clip, args.max_grad_norm)
                optimizer.step()
                lr_scheduler.step()
                optimizer.zero_grad()

                # Let's make sure we don't update any embedding weights besides the newly added token
                index_no_updates = torch.ones((len(tokenizer),), dtype=torch.bool)
                index_no_updates[min(placeholder_token_ids) : max(placeholder_token_ids) + 1] = False

                with torch.no_grad():
                    accelerator.unwrap_model(text_encoder).get_input_embeddings().weight[index_no_updates] = (
                        orig_embeds_params[index_no_updates]
                    )

            # Checks if the accelerator has performed an optimization step behind the scenes
            if accelerator.sync_gradients:
                progress_bar.update(1)
                global_step += 1
                accelerator.log({"train_loss": train_loss}, step=global_step)
                train_loss = 0.0

                if accelerator.is_main_process:
                    if global_step % args.checkpointing_steps == 0:
                        # _before_ saving state, check if this save would set us over the `checkpoints_total_limit`
                        if args.checkpoints_total_limit is not None:
                            checkpoints = os.listdir(args.output_dir)
                            checkpoints = [d for d in checkpoints if d.startswith("checkpoint")]
                            checkpoints = sorted(checkpoints, key=lambda x: int(x.split("-")[1]))

                            # before we save the new checkpoint, we need to have at _most_ `checkpoints_total_limit - 1` checkpoints
                            if len(checkpoints) >= args.checkpoints_total_limit:
                                num_to_remove = len(checkpoints) - args.checkpoints_total_limit + 1
                                removing_checkpoints = checkpoints[0:num_to_remove]

                                logger.info(
                                    f"{len(checkpoints)} checkpoints already exist, removing {len(removing_checkpoints)} checkpoints"
                                )
                                logger.info(f"removing checkpoints: {', '.join(removing_checkpoints)}")

                                for removing_checkpoint in removing_checkpoints:
                                    removing_checkpoint = os.path.join(args.output_dir, removing_checkpoint)
                                    shutil.rmtree(removing_checkpoint)

                        save_path = os.path.join(args.output_dir, f"checkpoint-{global_step}")
                        accelerator.save_state(save_path)
                        logger.info(f"Saved state to {save_path}")

                    if hasattr(args, "validation_data") and global_step % args.validation_steps == 0:
                        image_logs = log_validation(
                            tokenizer,
                            text_encoder,
                            brushnet,
                            args,
                            accelerator,
                            weight_dtype,
                            global_step,
                        )

            logs = {"step_loss": loss.detach().item(), "lr": lr_scheduler.get_last_lr()[0]}
            progress_bar.set_postfix(**logs)

            if global_step >= args.max_train_steps:
                break

    # Create the pipeline using using the trained modules and save it.
    accelerator.wait_for_everyone()
    if accelerator.is_main_process:
        brushnet = unwrap_model(brushnet)
        brushnet.save_pretrained(args.output_dir)

        # Run a final round of validation.
        image_logs = None
        if hasattr(args, "validation_data"):
            image_logs = log_validation(
                tokenizer,
                text_encoder,
                brushnet,
                args,
                accelerator,
                weight_dtype,
                global_step,
            )

        if args.push_to_hub:
            save_model_card(
                repo_id,
                image_logs=image_logs,
                base_model=args.pretrained_model_name_or_path,
                repo_folder=args.output_dir,
            )
            upload_folder(
                repo_id=repo_id,
                folder_path=args.output_dir,
                commit_message="End of training",
                ignore_patterns=["step_*", "epoch_*"],
            )

    accelerator.end_training()


if __name__ == "__main__":
    args = parse_args()
    main(args)