finetuning Stable Diffusion Super-Resolution model

[R2Bb1T]

Thanks for your works! Diffusers truly help me a lot to learn about diffusion models.
I've been trying to finetune the Stable Diffusion Super-Resolution model on my custom datasets. And I take the train_text_to_image_lora.py as example, but I am stucked on the problem that
torch.Size([2, 7, 128, 128]) torch.Size([2, 77, 1024]) Traceback (most recent call last): File "train_sr_lora.py", line 904, in <module> main() File "train_sr_lora.py", line 807, in main noise_pred = unet(torch.cat([lrimg,noisy_latents],dim=1), timesteps, encoder_hidden_states=prompt_embeds,class_labels=noise_level,return_dict=False).sample File "/opt/conda/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, **kwargs) File "/opt/conda/lib/python3.8/site-packages/diffusers/models/unet_2d_condition.py", line 724, in forward sample, res_samples = downsample_block( File "/opt/conda/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, **kwargs) File "/opt/conda/lib/python3.8/site-packages/diffusers/models/unet_2d_blocks.py", line 878, in forward hidden_states = attn( File "/opt/conda/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, **kwargs) File "/opt/conda/lib/python3.8/site-packages/diffusers/models/transformer_2d.py", line 265, in forward hidden_states = block( File "/opt/conda/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, **kwargs) File "/opt/conda/lib/python3.8/site-packages/diffusers/models/attention.py", line 143, in forward attn_output = self.attn1( File "/opt/conda/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, **kwargs) File "/opt/conda/lib/python3.8/site-packages/diffusers/models/attention_processor.py", line 299, in forward return self.processor( File "/opt/conda/lib/python3.8/site-packages/diffusers/models/attention_processor.py", line 531, in __call__ key = attn.to_k(encoder_hidden_states) + scale * self.to_k_lora(encoder_hidden_states) File "/opt/conda/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, **kwargs) File "/opt/conda/lib/python3.8/site-packages/diffusers/models/attention_processor.py", line 487, in forward down_hidden_states = self.down(hidden_states.to(dtype)) File "/opt/conda/lib/python3.8/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, **kwargs) File "/opt/conda/lib/python3.8/site-packages/torch/nn/modules/linear.py", line 94, in forward return F.linear(input, self.weight, self.bias) File "/opt/conda/lib/python3.8/site-packages/torch/nn/functional.py", line 1753, in linear return torch._C._nn.linear(input, weight, bias) RuntimeError: mat1 dim 1 must match mat2 dim 0
And I make sure the input image and prompt_embeds keep same with the code you provided in pipeline_stable_diffusion_upscale.py.
Both torch.Size([2, 7, 128, 128]) torch.Size([2, 77, 1024])
the training script is

# coding=utf-8
# Copyright 2023 The HuggingFace Inc. team. 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.
"""Fine-tuning script for Stable Diffusion for text2image with support for LoRA."""

import argparse
import logging
import math
import os
import random
from pathlib import Path

import datasets
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 packaging import version
from torchvision import transforms
from tqdm.auto import tqdm
from transformers import CLIPTextModel, CLIPTokenizer

import diffusers
from diffusers import AutoencoderKL, DDPMScheduler, DiffusionPipeline, UNet2DConditionModel
from diffusers.loaders import AttnProcsLayers
from diffusers.models.attention_processor import LoRAAttnProcessor
from diffusers.optimization import get_scheduler
from diffusers.utils import check_min_version, is_wandb_available,randn_tensor
from diffusers.utils.import_utils import is_xformers_available
from diffusers import StableDiffusionUpscalePipeline
from srdataset import AutoPairSRDataset, TextzoomDataset


# Will error if the minimal version of diffusers is not installed. Remove at your own risks.
check_min_version("0.17.0.dev0")

logger = get_logger(__name__, log_level="INFO")


def save_model_card(repo_id: str, images=None, base_model=str, dataset_name=str, repo_folder=None):
    img_str = ""
    for i, image in enumerate(images):
        image.save(os.path.join(repo_folder, f"image_{i}.png"))
        img_str += f"![img_{i}](./image_{i}.png)\n"

    yaml = f"""
---
license: creativeml-openrail-m
base_model: {base_model}
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
- lora
inference: true
---
    """
    model_card = f"""
# LoRA text2image fine-tuning - {repo_id}
These are LoRA adaption weights for {base_model}. The weights were fine-tuned on the {dataset_name} dataset. You can find some example images in the following. \n
{img_str}
"""
    with open(os.path.join(repo_folder, "README.md"), "w") as f:
        f.write(yaml + model_card)


def parse_args():
    parser = argparse.ArgumentParser(description="Simple example of a training script.")
    parser.add_argument(
        "--pretrained_model_name_or_path",
        type=str,
        default=None,
        required=True,
        help="Path to pretrained model or model identifier from huggingface.co/models.",
    )
    parser.add_argument(
        "--revision",
        type=str,
        default=None,
        required=False,
        help="Revision of pretrained model identifier from huggingface.co/models.",
    )
    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(
        "--train_data_dir",
        type=str,
        default=None,
        help=(
            "A folder containing the training data. Folder contents must follow the structure described in"
            " https://huggingface.co/docs/datasets/image_dataset#imagefolder. In particular, a `metadata.jsonl` file"
            " must exist to provide the captions for the images. Ignored if `dataset_name` is specified."
        ),
    )
    parser.add_argument(
        "--image_column", type=str, default="image", help="The column of the dataset containing an 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(
        "--validation_prompt", type=str, default=None, help="A prompt that is sampled during training for inference."
    )
    parser.add_argument(
        "--num_validation_images",
        type=int,
        default=4,
        help="Number of images that should be generated during validation with `validation_prompt`.",
    )
    parser.add_argument(
        "--validation_epochs",
        type=int,
        default=1,
        help=(
            "Run fine-tuning validation every X epochs. The validation process consists of running the prompt"
            " `args.validation_prompt` multiple times: `args.num_validation_images`."
        ),
    )
    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(
        "--output_dir",
        type=str,
        default="sd-model-finetuned-lora",
        help="The output directory where the model predictions and checkpoints will be written.",
    )
    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(
        "--center_crop",
        default=False,
        action="store_true",
        help=(
            "Whether to center crop the input images to the resolution. If not set, the images will be randomly"
            " cropped. The images will be resized to the resolution first before cropping."
        ),
    )
    parser.add_argument(
        "--random_flip",
        action="store_true",
        help="whether to randomly flip images horizontally",
    )
    parser.add_argument(
        "--train_batch_size", type=int, default=16, help="Batch size (per device) for the training dataloader."
    )
    parser.add_argument("--num_train_epochs", type=int, default=100)
    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(
        "--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=1e-4,
        help="Initial learning rate (after the potential warmup 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 warmup in the lr scheduler."
    )
    parser.add_argument(
        "--use_8bit_adam", action="store_true", help="Whether or not to use 8-bit Adam from bitsandbytes."
    )
    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(
        "--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(
        "--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(
        "--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("--local_rank", type=int, default=-1, help="For distributed training: local_rank")
    parser.add_argument(
        "--checkpointing_steps",
        type=int,
        default=500,
        help=(
            "Save a checkpoint of the training state every X updates. These checkpoints are only suitable for resuming"
            " training using `--resume_from_checkpoint`."
        ),
    )
    parser.add_argument(
        "--checkpoints_total_limit",
        type=int,
        default=None,
        help=(
            "Max number of checkpoints to store. Passed as `total_limit` to the `Accelerator` `ProjectConfiguration`."
            " See Accelerator::save_state https://huggingface.co/docs/accelerate/package_reference/accelerator#accelerate.Accelerator.save_state"
            " for more docs"
        ),
    )
    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(
        "--enable_xformers_memory_efficient_attention", action="store_true", help="Whether or not to use xformers."
    )
    parser.add_argument("--noise_offset", type=float, default=0, help="The scale of noise offset.")
    parser.add_argument("--paths",type=str,default = '/workspace/mnt/storage/yangzhuo/srpublicdata/textzoompng/train')
    parser.add_argument("--guidance_scale",type=float,default=1.0)
    parser.add_argument("--noise_level",type=int,default=20)
    args = parser.parse_args()
    env_local_rank = int(os.environ.get("LOCAL_RANK", -1))
    if env_local_rank != -1 and env_local_rank != args.local_rank:
        args.local_rank = env_local_rank

    # Sanity checks
    if args.dataset_name is None and args.train_data_dir is None:
        raise ValueError("Need either a dataset name or a training folder.")

    return args


DATASET_NAME_MAPPING = {
    "lambdalabs/pokemon-blip-captions": ("image", "text"),
}

def preprocess(image):
    if isinstance(image, torch.Tensor):
        return image
    elif isinstance(image, PIL.Image.Image):
        image = [image]

    if isinstance(image[0], PIL.Image.Image):
        w, h = image[0].size
        w, h = (x - x % 64 for x in (w, h))  # resize to integer multiple of 64

        image = [np.array(i.resize((w, h)))[None, :] for i in image]
        image = np.concatenate(image, axis=0)
        image = np.array(image).astype(np.float32) / 255.0
        image = image.transpose(0, 3, 1, 2)
        image = 2.0 * image - 1.0
        image = torch.from_numpy(image)
    elif isinstance(image[0], torch.Tensor):
        image = torch.cat(image, dim=0)
    return image

def main():
    args = parse_args()
    logging_dir = os.path.join(args.output_dir, args.logging_dir)

    accelerator_project_config = ProjectConfiguration(total_limit=args.checkpoints_total_limit)

    accelerator = Accelerator(
        gradient_accumulation_steps=args.gradient_accumulation_steps,
        mixed_precision=args.mixed_precision,
        log_with=args.report_to,
        logging_dir=logging_dir,
        project_config=accelerator_project_config,
    )
    if args.report_to == "wandb":
        if not is_wandb_available():
            raise ImportError("Make sure to install wandb if you want to use it for logging during training.")
        import wandb

    # 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:
        datasets.utils.logging.set_verbosity_warning()
        transformers.utils.logging.set_verbosity_warning()
        diffusers.utils.logging.set_verbosity_info()
    else:
        datasets.utils.logging.set_verbosity_error()
        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:
        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)

        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


    # Load scheduler, tokenizer and models.
    '''
    low_res_scheduler = DDPMScheduler.from_pretrained(args.pretrained_model_name_or_path, subfolder="scheduler")
    tokenizer = CLIPTokenizer.from_pretrained(
        args.pretrained_model_name_or_path, subfolder="tokenizer", revision=args.revision
    )
    text_encoder = CLIPTextModel.from_pretrained(
        args.pretrained_model_name_or_path, subfolder="text_encoder", revision=args.revision
    )
    vae = AutoencoderKL.from_pretrained(args.pretrained_model_name_or_path, subfolder="vae", revision=args.revision)
    unet = UNet2DConditionModel.from_pretrained(
        args.pretrained_model_name_or_path, subfolder="unet", revision=args.revision
    )
    '''
    model_id = "stabilityai/stable-diffusion-x4-upscaler"
    pipeline = StableDiffusionUpscalePipeline.from_pretrained(
        model_id, revision="fp16", torch_dtype=torch.float16
    )
    pipeline = pipeline.to("cuda")
    
    unet = pipeline.unet
    scheduler = pipeline.scheduler
    low_res_scheduler = pipeline.low_res_scheduler
    vae = pipeline.vae
    tokenizer = pipeline.tokenizer
    text_encoder = pipeline.text_encoder

    # freeze parameters of models to save more memory
    unet.requires_grad_(False)
    vae.requires_grad_(False)
    text_encoder.requires_grad_(False)

    # 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

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

    # now we will add new LoRA weights to the attention layers
    # It's important to realize here how many attention weights will be added and of which sizes
    # The sizes of the attention layers consist only of two different variables:
    # 1) - the "hidden_size", which is increased according to `unet.config.block_out_channels`.
    # 2) - the "cross attention size", which is set to `unet.config.cross_attention_dim`.

    # Let's first see how many attention processors we will have to set.
    # For Stable Diffusion, it should be equal to:
    # - down blocks (2x attention layers) * (2x transformer layers) * (3x down blocks) = 12
    # - mid blocks (2x attention layers) * (1x transformer layers) * (1x mid blocks) = 2
    # - up blocks (2x attention layers) * (3x transformer layers) * (3x down blocks) = 18
    # => 32 layers

    # Set correct lora layers
    lora_attn_procs = {}
    for name in unet.attn_processors.keys():
        cross_attention_dim = None if name.endswith("attn1.processor") else unet.config.cross_attention_dim
        if name.startswith("mid_block"):
            hidden_size = unet.config.block_out_channels[-1]
        elif name.startswith("up_blocks"):
            block_id = int(name[len("up_blocks.")])
            hidden_size = list(reversed(unet.config.block_out_channels))[block_id]
        elif name.startswith("down_blocks"):
            block_id = int(name[len("down_blocks.")])
            hidden_size = unet.config.block_out_channels[block_id]

        lora_attn_procs[name] = LoRAAttnProcessor(hidden_size=hidden_size, cross_attention_dim=cross_attention_dim)

    unet.set_attn_processor(lora_attn_procs)

    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()
        else:
            raise ValueError("xformers is not available. Make sure it is installed correctly")

    lora_layers = AttnProcsLayers(unet.attn_processors)

    # 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
        )

    # Initialize the optimizer
    if args.use_8bit_adam:
        try:
            import bitsandbytes as bnb
        except ImportError:
            raise ImportError(
                "Please install bitsandbytes to use 8-bit Adam. You can do so by running `pip install bitsandbytes`"
            )

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

    optimizer = optimizer_cls(
        lora_layers.parameters(),
        lr=args.learning_rate,
        betas=(args.adam_beta1, args.adam_beta2),
        weight_decay=args.adam_weight_decay,
        eps=args.adam_epsilon,
    )

    # Get the datasets: you can either provide your own training and evaluation files (see below)
    # or specify a Dataset from the hub (the dataset will be downloaded automatically from the datasets Hub).

    # In distributed training, the load_dataset function guarantees that only one local process can concurrently
    # download the dataset.
    '''
    if args.dataset_name is not None:
        # Downloading and loading a dataset from the hub.
        dataset = load_dataset(
            args.dataset_name,
            args.dataset_config_name,
            cache_dir=args.cache_dir,
        )
    else:
        data_files = {}
        if args.train_data_dir is not None:
            data_files["train"] = os.path.join(args.train_data_dir, "**")
        dataset = load_dataset(
            "imagefolder",
            data_files=data_files,
            cache_dir=args.cache_dir,
        )
        # See more about loading custom images at
        # https://huggingface.co/docs/datasets/v2.4.0/en/image_load#imagefolder

    # Preprocessing the datasets.
    # We need to tokenize inputs and targets.
    column_names = dataset["train"].column_names

    # 6. Get the column names for input/target.
    dataset_columns = DATASET_NAME_MAPPING.get(args.dataset_name, None)
    if args.image_column is None:
        image_column = dataset_columns[0] if dataset_columns is not None else column_names[0]
    else:
        image_column = args.image_column
        if image_column not in column_names:
            raise ValueError(
                f"--image_column' value '{args.image_column}' needs to be one of: {', '.join(column_names)}"
            )
    if args.caption_column is None:
        caption_column = dataset_columns[1] if dataset_columns is not None else column_names[1]
    else:
        caption_column = args.caption_column
        if caption_column not in column_names:
            raise ValueError(
                f"--caption_column' value '{args.caption_column}' needs to be one of: {', '.join(column_names)}"
            )

    # Preprocessing the datasets.
    # We need to tokenize input captions and transform the images.
    def tokenize_captions(examples, is_train=True):
        captions = []
        for caption in examples[caption_column]:
            if isinstance(caption, str):
                captions.append(caption)
            elif isinstance(caption, (list, np.ndarray)):
                # take a random caption if there are multiple
                captions.append(random.choice(caption) if is_train else caption[0])
            else:
                raise ValueError(
                    f"Caption column `{caption_column}` should contain either strings or lists of strings."
                )
        inputs = tokenizer(
            captions, max_length=tokenizer.model_max_length, padding="max_length", truncation=True, return_tensors="pt"
        )
        return inputs.input_ids

    # Preprocessing the datasets.
    train_transforms = transforms.Compose(
        [
            transforms.Resize(args.resolution, interpolation=transforms.InterpolationMode.BILINEAR),
            transforms.CenterCrop(args.resolution) if args.center_crop else transforms.RandomCrop(args.resolution),
            transforms.RandomHorizontalFlip() if args.random_flip else transforms.Lambda(lambda x: x),
            transforms.ToTensor(),
            transforms.Normalize([0.5], [0.5]),
        ]
    )

    def preprocess_train(examples):
        images = [image.convert("RGB") for image in examples[image_column]]
        examples["pixel_values"] = [train_transforms(image) for image in images]
        examples["input_ids"] = tokenize_captions(examples)
        return examples

    with accelerator.main_process_first():
        if args.max_train_samples is not None:
            dataset["train"] = dataset["train"].shuffle(seed=args.seed).select(range(args.max_train_samples))
        # Set the training transforms
        train_dataset = dataset["train"].with_transform(preprocess_train)

    def collate_fn(examples):
        pixel_values = torch.stack([example["pixel_values"] for example in examples])
        pixel_values = pixel_values.to(memory_format=torch.contiguous_format).float()
        input_ids = torch.stack([example["input_ids"] for example in examples])
        return {"pixel_values": pixel_values, "input_ids": input_ids}

    # DataLoaders creation:
    train_dataloader = torch.utils.data.DataLoader(
        train_dataset,
        shuffle=True,
        collate_fn=collate_fn,
        batch_size=args.train_batch_size,
        num_workers=args.dataloader_num_workers,
    )
    '''

    train_dataset = TextzoomDataset(args)
    #for content in train_dataset.labellst:
    #    tk = tokenizer(content, max_length=tokenizer.model_max_length, padding="max_length", truncation=True, return_tensors="pt")
    #    train_dataset.label.append(tk.input_ids[0])
    train_dataloader = torch.utils.data.DataLoader(train_dataset,batch_size=args.train_batch_size,shuffle=True,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 * args.gradient_accumulation_steps,
        num_training_steps=args.max_train_steps * args.gradient_accumulation_steps,
    )

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

    # 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:
        accelerator.init_trackers("text2image-fine-tune", config=vars(args))

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

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

            resume_global_step = global_step * args.gradient_accumulation_steps
            first_epoch = global_step // num_update_steps_per_epoch
            resume_step = resume_global_step % (num_update_steps_per_epoch * args.gradient_accumulation_steps)

    # Only show the progress bar once on each machine.
    progress_bar = tqdm(range(global_step, args.max_train_steps), disable=not accelerator.is_local_main_process)
    progress_bar.set_description("Steps")

    do_classifier_free_guidance = args.guidance_scale > 1.0

    for epoch in range(first_epoch, args.num_train_epochs):
        unet.train()#TODO
        train_loss = 0.0
        for step, batch in enumerate(train_dataloader):
            # Skip steps until we reach the resumed step
            lrimg, hrimg, prompt = batch#TODO
            batch_multiplier = 2 if do_classifier_free_guidance else 1
            hrimg = hrimg.to(device=accelerator.device, dtype=weight_dtype)
            lrimg = lrimg.to(device=accelerator.device, dtype=weight_dtype)

            if args.resume_from_checkpoint and epoch == first_epoch and step < resume_step:
                if step % args.gradient_accumulation_steps == 0:
                    progress_bar.update(1)
                continue

            with accelerator.accumulate(unet):
                # Convert images to latent space
                latents = vae.encode(hrimg).latent_dist.sample()
                latents = latents * vae.config.scaling_factor
                num_channels_latents = vae.config.latent_channels

                lrimg = torch.cat([lrimg] * batch_multiplier)
                latents = torch.cat([latents] * batch_multiplier)

                bsz = latents.shape[0]
                # Sample a random timestep for each image
                timesteps = torch.randint(0, low_res_scheduler.num_train_timesteps, (bsz,), device=accelerator.device)
                timesteps = timesteps.long()

                # Sample noise that we'll add to the latents
                noise_level = torch.tensor([args.noise_level], dtype=torch.long, device=accelerator.device)#TODO
                generator = torch.Generator(device=accelerator.device).manual_seed(args.seed)
                noise = randn_tensor(latents.shape, generator=generator, device=accelerator.device, dtype=latents.dtype)
                noisy_latents = low_res_scheduler.add_noise(latents, noise, noise_level) #可以直接得到加噪t次的xt图像

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

                # Get the text embedding for conditioning
                #batchsize=1
                #prompt_embeds = text_encoder(prompt)[0] #TODO
                
                prompt_embeds = pipeline._encode_prompt(
                    prompt[0],
                    accelerator.device,
                    num_images_per_prompt=1,
                    do_classifier_free_guidance=True
                )
                
                # Get the target for loss depending on the prediction type
                if low_res_scheduler.config.prediction_type == "epsilon":
                    target = noise
                elif low_res_scheduler.config.prediction_type == "v_prediction":
                    target = low_res_scheduler.get_velocity(latents, noise, timesteps)
                else:
                    raise ValueError(f"Unknown prediction type {low_res_scheduler.config.prediction_type}")
                
                num_channels_image = lrimg.shape[1]
                if num_channels_latents + num_channels_image != unet.config.in_channels:
                    raise ValueError(
                        f"Incorrect configuration settings! The config of `pipeline.unet`: {unet.config} expects"
                        f" {unet.config.in_channels} but received `num_channels_latents`: {num_channels_latents} +"
                        f" `num_channels_image`: {num_channels_image} "
                        f" = {num_channels_latents+num_channels_image}. Please verify the config of"
                        " `pipeline.unet` or your `image` input."
                    )
                #Prepare extra step kwargs. TODO: Logic should ideally just be moved out of the pipeline
                extra_step_kwargs = pipeline.prepare_extra_step_kwargs(generator, eta=0.0)

                print(torch.cat([lrimg,noisy_latents],dim=1).size(),prompt_embeds.size())
                # Predict the noise residual and compute loss
                noise_pred = unet(torch.cat([lrimg,noisy_latents],dim=1), timesteps, encoder_hidden_states=prompt_embeds,class_labels=noise_level,return_dict=False).sample

                # perform guidance
                if do_classifier_free_guidance:
                    noise_pred_uncond, noise_pred_text = noise_pred.chunk(2)
                    noise_pred = noise_pred_uncond + args.guidance_scale * (noise_pred_text - noise_pred_uncond)

                loss = F.l1_loss(noise_pred.float(), target.float(), reduction="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

                # Backpropagate
                accelerator.backward(loss)
                if accelerator.sync_gradients:
                    params_to_clip = lora_layers.parameters()
                    accelerator.clip_grad_norm_(params_to_clip, args.max_grad_norm)
                optimizer.step()
                lr_scheduler.step()
                optimizer.zero_grad()

            # 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 global_step % args.checkpointing_steps == 0:
                    if accelerator.is_main_process:
                        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}")

            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

    # Save the lora layers
    accelerator.wait_for_everyone()
    if accelerator.is_main_process:
        unet = unet.to(torch.float32)
        unet.save_attn_procs(args.output_dir)

        if args.push_to_hub:
            save_model_card(
                repo_id,
                images=images,
                base_model=args.pretrained_model_name_or_path,
                dataset_name=args.dataset_name,
                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_*"],
            )

    # Final inference
    # Load previous pipeline
    pipeline = DiffusionPipeline.from_pretrained(
        args.pretrained_model_name_or_path, revision=args.revision, torch_dtype=weight_dtype
    )
    pipeline = pipeline.to(accelerator.device)

    # load attention processors
    pipeline.unet.load_attn_procs(args.output_dir)

    # run inference
    generator = torch.Generator(device=accelerator.device).manual_seed(args.seed)
    images = []
    for _ in range(args.num_validation_images):
        images.append(pipeline(args.validation_prompt, num_inference_steps=30, generator=generator).images[0])

    if accelerator.is_main_process:
        for tracker in accelerator.trackers:
            if tracker.name == "tensorboard":
                np_images = np.stack([np.asarray(img) for img in images])
                tracker.writer.add_images("test", np_images, epoch, dataformats="NHWC")
            if tracker.name == "wandb":
                tracker.log(
                    {
                        "test": [
                            wandb.Image(image, caption=f"{i}: {args.validation_prompt}")
                            for i, image in enumerate(images)
                        ]
                    }
                )

    accelerator.end_training()


if __name__ == "__main__":
    main()

The dataset code is

class TextzoomDataset(Dataset):
    def __init__(self,args):#TODO:mode
        self.args = args
        self.lrlst = [args.paths+'/lr/{}.png'.format(i) for i in range(17368)]
        self.hrlst = [args.paths+'/hr/{}.png'.format(i) for i in range(17368)]
        self.label = open(args.paths+'/label.txt','r').readlines()
        #self.label = []
        self.transforms = transforms.Compose([
            transforms.ToTensor(),
            transforms.Normalize([0.5],[0.5])]
        )
    
    def __len__(self):
        return len(self.hrlst)
    
    def __getitem__(self, index):
        lrimg,hrimg,label = Image.open(self.lrlst[index]),Image.open(self.hrlst[index]),self.label[index]
        lw,lh = lrimg.size
        hw,hh = hrimg.size

        if lw<3*lh//2:
            lrimg.resize((lw,lw),Image.BILINEAR)
            hrimg.resize((hw,hw),Image.BILINEAR)
        elif lw<5*lh//2:
            lrimg.resize((lw,lw//2),Image.BILINEAR)
            hrimg.resize((hw,hw//2),Image.BILINEAR)
        else:
            lrimg.resize((lw,lw//3),Image.BILINEAR)
            hrimg.resize((hw,hw//3),Image.BILINEAR)
        lrimg = self.transforms(lrimg)
        hrimg = self.transforms(hrimg) 
        if lw<5*lh//2 and lw>=3*lh//2:
            lrimg = torch.cat([lrimg,lrimg],dim=1)
            hrimg = torch.cat([hrimg,hrimg],dim=1)
        elif lw>=5*lh//2:
            lrimg = torch.cat([lrimg,lrimg,lrimg],dim=1)
            hrimg = torch.cat([hrimg,hrimg,hrimg],dim=1)
        lrimg=F.interpolate(lrimg.unsqueeze(0),size=(128,128),mode='bilinear').squeeze(0)
        hrimg=F.interpolate(hrimg.unsqueeze(0),size=(512,512),mode='bilinear').squeeze(0)
            

        return lrimg, hrimg, label[:-1]

Besides I searched about relative finetuning issue on github and it seems still no solid progress yet. And I'm a new beginner to diffusion models, is there community or something to communicate ?

[rupertmenneer]

Looks like the issue you have here is probably down to tensor dimension mismatching, see the description in the error you have attached:

torch.Size([2, 7, 128, 128]) torch.Size([2, 77, 1024]) Traceback (most recent call last): File "train_sr_lora.py", line 904, in <module> main() File "train_sr_lora.py", line 807, in main noise_pred = unet(torch.cat([lrimg,noisy_latents],dim=1)

torch.cat([tensor_1, tensor_2], dim=1) must be able to be concatenated along dimension 1, but they don't share the same dimensions ([128, 128] and [1024])

Looks like the actual example train_text_to_image_lora.py they are passing noisy latents to their model whereas you are trying to pass both noisy latents and some form of a low resolution image, but the dimensions don't match in the torch.cat function, so you get an error!

Since you are fine-tuning a super res model, I would first start with your low res and high res images and making sure they match they should be something like this:

Batch * Channels * Height * Width
and
Batch * Channels * (Height*scale) * (Width*scale)

It looks like your TextzoomDataset has some odd code! I don't know all the ins and outs of what you're trying but I would start there, then make sure that if you are adding your low res image as an additional condition then also pass them into latent space before concatenation so they share the same dimensions? That would be my guess at what is going on here! Hope that is helpful :)

[R2Bb1T]

@rupertmenneer Thanks for replying!
I had checked the 'lrimg' and 'noisy_latents' and I am sure they are in the same size, and the size mismatch you mentioned is between the catted tensor (2,7,128,128) which did not report error and the prompt_embeds(2,77,1024). And I've checked the original pipeline and its corresponding inputs size are both (2,7,128,128) and (2,77,1024). So it is really confusing that it reports size mismatching.

Update: I just found that the noisy_latents are all nan value, and I am pretty sure this is the reason.

[R2Bb1T]

Update: Now I am stucked on the reason output latents of vae are all nan. Is it because the distribution of textzoom dataset unseen for it? Besides, my process to the textzoom dataset is to stack the same image to make its width and height more close as the most images in the textzoom dataset got larger width than height.
I am checking if I preprocess them right.

[R2Bb1T]

Preprocess seems fine, the dataset address is https://github.com/WenjiaWang0312/TextZoom.

[sayakpaul]

I think it might have to do with the hyperparameters. Maybe start by trying to overfit a single batch of data?

Since this issue is not related to our codebase, I am going to close the issue and welcome you to post it on the forums: https://discuss.huggingface.co/c/discussion-related-to-httpsgithubcomhuggingfacediffusers/63