|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
import argparse |
|
import gc |
|
import hashlib |
|
import itertools |
|
import logging |
|
import math |
|
import os |
|
import shutil |
|
import warnings |
|
from pathlib import Path |
|
from typing import Dict |
|
|
|
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 PIL import Image |
|
from PIL.ImageOps import exif_transpose |
|
from torch.utils.data import Dataset |
|
from torchvision import transforms |
|
from tqdm.auto import tqdm |
|
from transformers import AutoTokenizer, PretrainedConfig |
|
|
|
import diffusers |
|
from diffusers import ( |
|
AutoencoderKL, |
|
DDPMScheduler, |
|
DPMSolverMultistepScheduler, |
|
StableDiffusionXLPipeline, |
|
UNet2DConditionModel, |
|
) |
|
from diffusers.loaders import LoraLoaderMixin, text_encoder_lora_state_dict |
|
from diffusers.models.attention_processor import LoRAAttnProcessor, LoRAAttnProcessor2_0 |
|
from diffusers.optimization import get_scheduler |
|
from diffusers.utils import check_min_version, is_wandb_available |
|
from diffusers.utils.import_utils import is_xformers_available |
|
|
|
|
|
|
|
check_min_version("0.21.0.dev0") |
|
|
|
logger = get_logger(__name__) |
|
|
|
|
|
def save_model_card( |
|
repo_id: str, images=None, base_model=str, train_text_encoder=False, prompt=str, repo_folder=None, vae_path=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: openrail++ |
|
base_model: {base_model} |
|
instance_prompt: {prompt} |
|
tags: |
|
- stable-diffusion-xl |
|
- stable-diffusion-xl-diffusers |
|
- text-to-image |
|
- diffusers |
|
- lora |
|
inference: true |
|
--- |
|
""" |
|
model_card = f""" |
|
# LoRA DreamBooth - {repo_id} |
|
|
|
These are LoRA adaption weights for {base_model}. The weights were trained on {prompt} using [DreamBooth](https://dreambooth.github.io/). You can find some example images in the following. \n |
|
{img_str} |
|
|
|
LoRA for the text encoder was enabled: {train_text_encoder}. |
|
|
|
Special VAE used for training: {vae_path}. |
|
""" |
|
with open(os.path.join(repo_folder, "README.md"), "w") as f: |
|
f.write(yaml + model_card) |
|
|
|
|
|
def import_model_class_from_model_name_or_path( |
|
pretrained_model_name_or_path: str, revision: str, subfolder: str = "text_encoder" |
|
): |
|
text_encoder_config = PretrainedConfig.from_pretrained( |
|
pretrained_model_name_or_path, subfolder=subfolder, revision=revision |
|
) |
|
model_class = text_encoder_config.architectures[0] |
|
|
|
if model_class == "CLIPTextModel": |
|
from transformers import CLIPTextModel |
|
|
|
return CLIPTextModel |
|
elif model_class == "CLIPTextModelWithProjection": |
|
from transformers import CLIPTextModelWithProjection |
|
|
|
return CLIPTextModelWithProjection |
|
else: |
|
raise ValueError(f"{model_class} is not supported.") |
|
|
|
|
|
def parse_args(input_args=None): |
|
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( |
|
"--pretrained_vae_model_name_or_path", |
|
type=str, |
|
default=None, |
|
help="Path to pretrained VAE model with better numerical stability. More details: https://github.com/huggingface/diffusers/pull/4038.", |
|
) |
|
parser.add_argument( |
|
"--revision", |
|
type=str, |
|
default=None, |
|
required=False, |
|
help="Revision of pretrained model identifier from huggingface.co/models.", |
|
) |
|
parser.add_argument( |
|
"--instance_data_dir", |
|
type=str, |
|
default=None, |
|
required=True, |
|
help="A folder containing the training data of instance images.", |
|
) |
|
parser.add_argument( |
|
"--class_data_dir", |
|
type=str, |
|
default=None, |
|
required=False, |
|
help="A folder containing the training data of class images.", |
|
) |
|
parser.add_argument( |
|
"--instance_prompt", |
|
type=str, |
|
default=None, |
|
required=True, |
|
help="The prompt with identifier specifying the instance", |
|
) |
|
parser.add_argument( |
|
"--class_prompt", |
|
type=str, |
|
default=None, |
|
help="The prompt to specify images in the same class as provided instance images.", |
|
) |
|
parser.add_argument( |
|
"--validation_prompt", |
|
type=str, |
|
default=None, |
|
help="A prompt that is used during validation to verify that the model is learning.", |
|
) |
|
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=50, |
|
help=( |
|
"Run dreambooth validation every X epochs. Dreambooth validation consists of running the prompt" |
|
" `args.validation_prompt` multiple times: `args.num_validation_images`." |
|
), |
|
) |
|
parser.add_argument( |
|
"--with_prior_preservation", |
|
default=False, |
|
action="store_true", |
|
help="Flag to add prior preservation loss.", |
|
) |
|
parser.add_argument("--prior_loss_weight", type=float, default=1.0, help="The weight of prior preservation loss.") |
|
parser.add_argument( |
|
"--num_class_images", |
|
type=int, |
|
default=100, |
|
help=( |
|
"Minimal class images for prior preservation loss. If there are not enough images already present in" |
|
" class_data_dir, additional images will be sampled with class_prompt." |
|
), |
|
) |
|
parser.add_argument( |
|
"--output_dir", |
|
type=str, |
|
default="lora-dreambooth-model", |
|
help="The output directory where the model predictions and checkpoints will be written.", |
|
) |
|
parser.add_argument("--seed", type=int, default=None, help="A seed for reproducible training.") |
|
parser.add_argument( |
|
"--resolution", |
|
type=int, |
|
default=1024, |
|
help=( |
|
"The resolution for input images, all the images in the train/validation dataset will be resized to this" |
|
" resolution" |
|
), |
|
) |
|
parser.add_argument( |
|
"--crops_coords_top_left_h", |
|
type=int, |
|
default=0, |
|
help=("Coordinate for (the height) to be included in the crop coordinate embeddings needed by SDXL UNet."), |
|
) |
|
parser.add_argument( |
|
"--crops_coords_top_left_w", |
|
type=int, |
|
default=0, |
|
help=("Coordinate for (the height) to be included in the crop coordinate embeddings needed by SDXL UNet."), |
|
) |
|
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( |
|
"--train_text_encoder", |
|
action="store_true", |
|
help="Whether to train the text encoder. If set, the text encoder should be float32 precision.", |
|
) |
|
parser.add_argument( |
|
"--train_batch_size", type=int, default=4, help="Batch size (per device) for the training dataloader." |
|
) |
|
parser.add_argument( |
|
"--sample_batch_size", type=int, default=4, help="Batch size (per device) for sampling images." |
|
) |
|
parser.add_argument("--num_train_epochs", type=int, default=1) |
|
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. These checkpoints can be used both as final" |
|
" checkpoints in case they are better than the last checkpoint, and are also 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."), |
|
) |
|
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-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( |
|
"--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( |
|
"--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( |
|
"--use_8bit_adam", action="store_true", help="Whether or not to use 8-bit Adam from bitsandbytes." |
|
) |
|
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( |
|
"--prior_generation_precision", |
|
type=str, |
|
default=None, |
|
choices=["no", "fp32", "fp16", "bf16"], |
|
help=( |
|
"Choose prior generation precision between fp32, fp16 and bf16 (bfloat16). Bf16 requires PyTorch >=" |
|
" 1.10.and an Nvidia Ampere GPU. Default to fp16 if a GPU is available else fp32." |
|
), |
|
) |
|
parser.add_argument("--local_rank", type=int, default=-1, help="For distributed training: local_rank") |
|
parser.add_argument( |
|
"--enable_xformers_memory_efficient_attention", action="store_true", help="Whether or not to use xformers." |
|
) |
|
parser.add_argument( |
|
"--rank", |
|
type=int, |
|
default=4, |
|
help=("The dimension of the LoRA update matrices."), |
|
) |
|
|
|
if input_args is not None: |
|
args = parser.parse_args(input_args) |
|
else: |
|
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 |
|
|
|
if args.with_prior_preservation: |
|
if args.class_data_dir is None: |
|
raise ValueError("You must specify a data directory for class images.") |
|
if args.class_prompt is None: |
|
raise ValueError("You must specify prompt for class images.") |
|
else: |
|
|
|
if args.class_data_dir is not None: |
|
warnings.warn("You need not use --class_data_dir without --with_prior_preservation.") |
|
if args.class_prompt is not None: |
|
warnings.warn("You need not use --class_prompt without --with_prior_preservation.") |
|
|
|
return args |
|
|
|
|
|
class DreamBoothDataset(Dataset): |
|
""" |
|
A dataset to prepare the instance and class images with the prompts for fine-tuning the model. |
|
It pre-processes the images. |
|
""" |
|
|
|
def __init__( |
|
self, |
|
instance_data_root, |
|
class_data_root=None, |
|
class_num=None, |
|
size=1024, |
|
center_crop=False, |
|
): |
|
self.size = size |
|
self.center_crop = center_crop |
|
|
|
self.instance_data_root = Path(instance_data_root) |
|
if not self.instance_data_root.exists(): |
|
raise ValueError("Instance images root doesn't exists.") |
|
|
|
self.instance_images_path = list(Path(instance_data_root).iterdir()) |
|
self.num_instance_images = len(self.instance_images_path) |
|
self._length = self.num_instance_images |
|
|
|
if class_data_root is not None: |
|
self.class_data_root = Path(class_data_root) |
|
self.class_data_root.mkdir(parents=True, exist_ok=True) |
|
self.class_images_path = list(self.class_data_root.iterdir()) |
|
if class_num is not None: |
|
self.num_class_images = min(len(self.class_images_path), class_num) |
|
else: |
|
self.num_class_images = len(self.class_images_path) |
|
self._length = max(self.num_class_images, self.num_instance_images) |
|
else: |
|
self.class_data_root = None |
|
|
|
self.image_transforms = transforms.Compose( |
|
[ |
|
transforms.Resize(size, interpolation=transforms.InterpolationMode.BILINEAR), |
|
transforms.CenterCrop(size) if center_crop else transforms.RandomCrop(size), |
|
transforms.ToTensor(), |
|
transforms.Normalize([0.5], [0.5]), |
|
] |
|
) |
|
|
|
def __len__(self): |
|
return self._length |
|
|
|
def __getitem__(self, index): |
|
example = {} |
|
instance_image = Image.open(self.instance_images_path[index % self.num_instance_images]) |
|
instance_image = exif_transpose(instance_image) |
|
|
|
if not instance_image.mode == "RGB": |
|
instance_image = instance_image.convert("RGB") |
|
example["instance_images"] = self.image_transforms(instance_image) |
|
|
|
if self.class_data_root: |
|
class_image = Image.open(self.class_images_path[index % self.num_class_images]) |
|
class_image = exif_transpose(class_image) |
|
|
|
if not class_image.mode == "RGB": |
|
class_image = class_image.convert("RGB") |
|
example["class_images"] = self.image_transforms(class_image) |
|
|
|
return example |
|
|
|
|
|
def collate_fn(examples, with_prior_preservation=False): |
|
pixel_values = [example["instance_images"] for example in examples] |
|
|
|
|
|
|
|
if with_prior_preservation: |
|
pixel_values += [example["class_images"] for example in examples] |
|
|
|
pixel_values = torch.stack(pixel_values) |
|
pixel_values = pixel_values.to(memory_format=torch.contiguous_format).float() |
|
|
|
batch = {"pixel_values": pixel_values} |
|
return batch |
|
|
|
|
|
class PromptDataset(Dataset): |
|
"A simple dataset to prepare the prompts to generate class images on multiple GPUs." |
|
|
|
def __init__(self, prompt, num_samples): |
|
self.prompt = prompt |
|
self.num_samples = num_samples |
|
|
|
def __len__(self): |
|
return self.num_samples |
|
|
|
def __getitem__(self, index): |
|
example = {} |
|
example["prompt"] = self.prompt |
|
example["index"] = index |
|
return example |
|
|
|
|
|
def tokenize_prompt(tokenizer, prompt): |
|
text_inputs = tokenizer( |
|
prompt, |
|
padding="max_length", |
|
max_length=tokenizer.model_max_length, |
|
truncation=True, |
|
return_tensors="pt", |
|
) |
|
text_input_ids = text_inputs.input_ids |
|
return text_input_ids |
|
|
|
|
|
|
|
def encode_prompt(text_encoders, tokenizers, prompt, text_input_ids_list=None): |
|
prompt_embeds_list = [] |
|
|
|
for i, text_encoder in enumerate(text_encoders): |
|
if tokenizers is not None: |
|
tokenizer = tokenizers[i] |
|
text_input_ids = tokenize_prompt(tokenizer, prompt) |
|
else: |
|
assert text_input_ids_list is not None |
|
text_input_ids = text_input_ids_list[i] |
|
|
|
prompt_embeds = text_encoder( |
|
text_input_ids.to(text_encoder.device), |
|
output_hidden_states=True, |
|
) |
|
|
|
|
|
pooled_prompt_embeds = prompt_embeds[0] |
|
prompt_embeds = prompt_embeds.hidden_states[-2] |
|
bs_embed, seq_len, _ = prompt_embeds.shape |
|
prompt_embeds = prompt_embeds.view(bs_embed, seq_len, -1) |
|
prompt_embeds_list.append(prompt_embeds) |
|
|
|
prompt_embeds = torch.concat(prompt_embeds_list, dim=-1) |
|
pooled_prompt_embeds = pooled_prompt_embeds.view(bs_embed, -1) |
|
return prompt_embeds, pooled_prompt_embeds |
|
|
|
|
|
def unet_attn_processors_state_dict(unet) -> Dict[str, torch.tensor]: |
|
""" |
|
Returns: |
|
a state dict containing just the attention processor parameters. |
|
""" |
|
attn_processors = unet.attn_processors |
|
|
|
attn_processors_state_dict = {} |
|
|
|
for attn_processor_key, attn_processor in attn_processors.items(): |
|
for parameter_key, parameter in attn_processor.state_dict().items(): |
|
attn_processors_state_dict[f"{attn_processor_key}.{parameter_key}"] = parameter |
|
|
|
return attn_processors_state_dict |
|
|
|
|
|
def main(args): |
|
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, |
|
) |
|
|
|
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 |
|
|
|
|
|
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 args.seed is not None: |
|
set_seed(args.seed) |
|
|
|
|
|
if args.with_prior_preservation: |
|
class_images_dir = Path(args.class_data_dir) |
|
if not class_images_dir.exists(): |
|
class_images_dir.mkdir(parents=True) |
|
cur_class_images = len(list(class_images_dir.iterdir())) |
|
|
|
if cur_class_images < args.num_class_images: |
|
torch_dtype = torch.float16 if accelerator.device.type == "cuda" else torch.float32 |
|
if args.prior_generation_precision == "fp32": |
|
torch_dtype = torch.float32 |
|
elif args.prior_generation_precision == "fp16": |
|
torch_dtype = torch.float16 |
|
elif args.prior_generation_precision == "bf16": |
|
torch_dtype = torch.bfloat16 |
|
pipeline = StableDiffusionXLPipeline.from_pretrained( |
|
args.pretrained_model_name_or_path, |
|
torch_dtype=torch_dtype, |
|
revision=args.revision, |
|
) |
|
pipeline.set_progress_bar_config(disable=True) |
|
|
|
num_new_images = args.num_class_images - cur_class_images |
|
logger.info(f"Number of class images to sample: {num_new_images}.") |
|
|
|
sample_dataset = PromptDataset(args.class_prompt, num_new_images) |
|
sample_dataloader = torch.utils.data.DataLoader(sample_dataset, batch_size=args.sample_batch_size) |
|
|
|
sample_dataloader = accelerator.prepare(sample_dataloader) |
|
pipeline.to(accelerator.device) |
|
|
|
for example in tqdm( |
|
sample_dataloader, desc="Generating class images", disable=not accelerator.is_local_main_process |
|
): |
|
images = pipeline(example["prompt"]).images |
|
|
|
for i, image in enumerate(images): |
|
hash_image = hashlib.sha1(image.tobytes()).hexdigest() |
|
image_filename = class_images_dir / f"{example['index'][i] + cur_class_images}-{hash_image}.jpg" |
|
image.save(image_filename) |
|
|
|
del pipeline |
|
if torch.cuda.is_available(): |
|
torch.cuda.empty_cache() |
|
|
|
|
|
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 |
|
|
|
|
|
tokenizer_one = AutoTokenizer.from_pretrained( |
|
args.pretrained_model_name_or_path, subfolder="tokenizer", revision=args.revision, use_fast=False |
|
) |
|
tokenizer_two = AutoTokenizer.from_pretrained( |
|
args.pretrained_model_name_or_path, subfolder="tokenizer_2", revision=args.revision, use_fast=False |
|
) |
|
|
|
|
|
text_encoder_cls_one = import_model_class_from_model_name_or_path( |
|
args.pretrained_model_name_or_path, args.revision |
|
) |
|
text_encoder_cls_two = import_model_class_from_model_name_or_path( |
|
args.pretrained_model_name_or_path, args.revision, subfolder="text_encoder_2" |
|
) |
|
|
|
|
|
noise_scheduler = DDPMScheduler.from_pretrained(args.pretrained_model_name_or_path, subfolder="scheduler") |
|
text_encoder_one = text_encoder_cls_one.from_pretrained( |
|
args.pretrained_model_name_or_path, subfolder="text_encoder", revision=args.revision |
|
) |
|
text_encoder_two = text_encoder_cls_two.from_pretrained( |
|
args.pretrained_model_name_or_path, subfolder="text_encoder_2", revision=args.revision |
|
) |
|
vae_path = ( |
|
args.pretrained_model_name_or_path |
|
if args.pretrained_vae_model_name_or_path is None |
|
else args.pretrained_vae_model_name_or_path |
|
) |
|
vae = AutoencoderKL.from_pretrained( |
|
vae_path, subfolder="vae" if args.pretrained_vae_model_name_or_path is None else None, revision=args.revision |
|
) |
|
unet = UNet2DConditionModel.from_pretrained( |
|
args.pretrained_model_name_or_path, subfolder="unet", revision=args.revision |
|
) |
|
|
|
|
|
vae.requires_grad_(False) |
|
text_encoder_one.requires_grad_(False) |
|
text_encoder_two.requires_grad_(False) |
|
unet.requires_grad_(False) |
|
|
|
|
|
|
|
weight_dtype = torch.float32 |
|
if accelerator.mixed_precision == "fp16": |
|
weight_dtype = torch.float16 |
|
elif accelerator.mixed_precision == "bf16": |
|
weight_dtype = torch.bfloat16 |
|
|
|
|
|
unet.to(accelerator.device, dtype=weight_dtype) |
|
|
|
|
|
vae.to(accelerator.device, dtype=torch.float32) |
|
|
|
text_encoder_one.to(accelerator.device, dtype=weight_dtype) |
|
text_encoder_two.to(accelerator.device, dtype=weight_dtype) |
|
|
|
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") |
|
|
|
if args.gradient_checkpointing: |
|
unet.enable_gradient_checkpointing() |
|
if args.train_text_encoder: |
|
text_encoder_one.gradient_checkpointing_enable() |
|
text_encoder_two.gradient_checkpointing_enable() |
|
|
|
|
|
|
|
unet_lora_attn_procs = {} |
|
unet_lora_parameters = [] |
|
for name, attn_processor in unet.attn_processors.items(): |
|
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_processor_class = ( |
|
LoRAAttnProcessor2_0 if hasattr(F, "scaled_dot_product_attention") else LoRAAttnProcessor |
|
) |
|
module = lora_attn_processor_class( |
|
hidden_size=hidden_size, cross_attention_dim=cross_attention_dim, rank=args.rank |
|
) |
|
unet_lora_attn_procs[name] = module |
|
unet_lora_parameters.extend(module.parameters()) |
|
|
|
unet.set_attn_processor(unet_lora_attn_procs) |
|
|
|
|
|
|
|
if args.train_text_encoder: |
|
|
|
text_lora_parameters_one = LoraLoaderMixin._modify_text_encoder( |
|
text_encoder_one, dtype=torch.float32, rank=args.rank |
|
) |
|
text_lora_parameters_two = LoraLoaderMixin._modify_text_encoder( |
|
text_encoder_two, dtype=torch.float32, rank=args.rank |
|
) |
|
|
|
|
|
def save_model_hook(models, weights, output_dir): |
|
if accelerator.is_main_process: |
|
|
|
|
|
unet_lora_layers_to_save = None |
|
text_encoder_one_lora_layers_to_save = None |
|
text_encoder_two_lora_layers_to_save = None |
|
|
|
for model in models: |
|
if isinstance(model, type(accelerator.unwrap_model(unet))): |
|
unet_lora_layers_to_save = unet_attn_processors_state_dict(model) |
|
elif isinstance(model, type(accelerator.unwrap_model(text_encoder_one))): |
|
text_encoder_one_lora_layers_to_save = text_encoder_lora_state_dict(model) |
|
elif isinstance(model, type(accelerator.unwrap_model(text_encoder_two))): |
|
text_encoder_two_lora_layers_to_save = text_encoder_lora_state_dict(model) |
|
else: |
|
raise ValueError(f"unexpected save model: {model.__class__}") |
|
|
|
|
|
weights.pop() |
|
|
|
StableDiffusionXLPipeline.save_lora_weights( |
|
output_dir, |
|
unet_lora_layers=unet_lora_layers_to_save, |
|
text_encoder_lora_layers=text_encoder_one_lora_layers_to_save, |
|
text_encoder_2_lora_layers=text_encoder_two_lora_layers_to_save, |
|
) |
|
|
|
def load_model_hook(models, input_dir): |
|
unet_ = None |
|
text_encoder_one_ = None |
|
text_encoder_two_ = None |
|
|
|
while len(models) > 0: |
|
model = models.pop() |
|
|
|
if isinstance(model, type(accelerator.unwrap_model(unet))): |
|
unet_ = model |
|
elif isinstance(model, type(accelerator.unwrap_model(text_encoder_one))): |
|
text_encoder_one_ = model |
|
elif isinstance(model, type(accelerator.unwrap_model(text_encoder_two))): |
|
text_encoder_two_ = model |
|
else: |
|
raise ValueError(f"unexpected save model: {model.__class__}") |
|
|
|
lora_state_dict, network_alphas = LoraLoaderMixin.lora_state_dict(input_dir) |
|
LoraLoaderMixin.load_lora_into_unet(lora_state_dict, network_alphas=network_alphas, unet=unet_) |
|
|
|
text_encoder_state_dict = {k: v for k, v in lora_state_dict.items() if "text_encoder." in k} |
|
LoraLoaderMixin.load_lora_into_text_encoder( |
|
text_encoder_state_dict, network_alphas=network_alphas, text_encoder=text_encoder_one_ |
|
) |
|
|
|
text_encoder_2_state_dict = {k: v for k, v in lora_state_dict.items() if "text_encoder_2." in k} |
|
LoraLoaderMixin.load_lora_into_text_encoder( |
|
text_encoder_2_state_dict, network_alphas=network_alphas, text_encoder=text_encoder_two_ |
|
) |
|
|
|
accelerator.register_save_state_pre_hook(save_model_hook) |
|
accelerator.register_load_state_pre_hook(load_model_hook) |
|
|
|
|
|
|
|
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 |
|
) |
|
|
|
|
|
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 |
|
|
|
|
|
params_to_optimize = ( |
|
itertools.chain(unet_lora_parameters, text_lora_parameters_one, text_lora_parameters_two) |
|
if args.train_text_encoder |
|
else unet_lora_parameters |
|
) |
|
optimizer = optimizer_class( |
|
params_to_optimize, |
|
lr=args.learning_rate, |
|
betas=(args.adam_beta1, args.adam_beta2), |
|
weight_decay=args.adam_weight_decay, |
|
eps=args.adam_epsilon, |
|
) |
|
|
|
|
|
|
|
|
|
|
|
|
|
def compute_time_ids(): |
|
|
|
original_size = (args.resolution, args.resolution) |
|
target_size = (args.resolution, args.resolution) |
|
crops_coords_top_left = (args.crops_coords_top_left_h, args.crops_coords_top_left_w) |
|
add_time_ids = list(original_size + crops_coords_top_left + target_size) |
|
add_time_ids = torch.tensor([add_time_ids]) |
|
add_time_ids = add_time_ids.to(accelerator.device, dtype=weight_dtype) |
|
return add_time_ids |
|
|
|
if not args.train_text_encoder: |
|
tokenizers = [tokenizer_one, tokenizer_two] |
|
text_encoders = [text_encoder_one, text_encoder_two] |
|
|
|
def compute_text_embeddings(prompt, text_encoders, tokenizers): |
|
with torch.no_grad(): |
|
prompt_embeds, pooled_prompt_embeds = encode_prompt(text_encoders, tokenizers, prompt) |
|
prompt_embeds = prompt_embeds.to(accelerator.device) |
|
pooled_prompt_embeds = pooled_prompt_embeds.to(accelerator.device) |
|
return prompt_embeds, pooled_prompt_embeds |
|
|
|
|
|
instance_time_ids = compute_time_ids() |
|
if not args.train_text_encoder: |
|
instance_prompt_hidden_states, instance_pooled_prompt_embeds = compute_text_embeddings( |
|
args.instance_prompt, text_encoders, tokenizers |
|
) |
|
|
|
|
|
if args.with_prior_preservation: |
|
class_time_ids = compute_time_ids() |
|
if not args.train_text_encoder: |
|
class_prompt_hidden_states, class_pooled_prompt_embeds = compute_text_embeddings( |
|
args.class_prompt, text_encoders, tokenizers |
|
) |
|
|
|
|
|
if not args.train_text_encoder: |
|
del tokenizers, text_encoders |
|
gc.collect() |
|
torch.cuda.empty_cache() |
|
|
|
|
|
|
|
add_time_ids = instance_time_ids |
|
if args.with_prior_preservation: |
|
add_time_ids = torch.cat([add_time_ids, class_time_ids], dim=0) |
|
|
|
if not args.train_text_encoder: |
|
prompt_embeds = instance_prompt_hidden_states |
|
unet_add_text_embeds = instance_pooled_prompt_embeds |
|
if args.with_prior_preservation: |
|
prompt_embeds = torch.cat([prompt_embeds, class_prompt_hidden_states], dim=0) |
|
unet_add_text_embeds = torch.cat([unet_add_text_embeds, class_pooled_prompt_embeds], dim=0) |
|
else: |
|
tokens_one = tokenize_prompt(tokenizer_one, args.instance_prompt) |
|
tokens_two = tokenize_prompt(tokenizer_two, args.instance_prompt) |
|
if args.with_prior_preservation: |
|
class_tokens_one = tokenize_prompt(tokenizer_one, args.class_prompt) |
|
class_tokens_two = tokenize_prompt(tokenizer_two, args.class_prompt) |
|
tokens_one = torch.cat([tokens_one, class_tokens_one], dim=0) |
|
tokens_two = torch.cat([tokens_two, class_tokens_two], dim=0) |
|
|
|
|
|
train_dataset = DreamBoothDataset( |
|
instance_data_root=args.instance_data_dir, |
|
class_data_root=args.class_data_dir if args.with_prior_preservation else None, |
|
class_num=args.num_class_images, |
|
size=args.resolution, |
|
center_crop=args.center_crop, |
|
) |
|
|
|
train_dataloader = torch.utils.data.DataLoader( |
|
train_dataset, |
|
batch_size=args.train_batch_size, |
|
shuffle=True, |
|
collate_fn=lambda examples: collate_fn(examples, args.with_prior_preservation), |
|
num_workers=args.dataloader_num_workers, |
|
) |
|
|
|
|
|
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, |
|
) |
|
|
|
|
|
if args.train_text_encoder: |
|
unet, text_encoder_one, text_encoder_two, optimizer, train_dataloader, lr_scheduler = accelerator.prepare( |
|
unet, text_encoder_one, text_encoder_two, optimizer, train_dataloader, lr_scheduler |
|
) |
|
else: |
|
unet, optimizer, train_dataloader, lr_scheduler = accelerator.prepare( |
|
unet, optimizer, train_dataloader, lr_scheduler |
|
) |
|
|
|
|
|
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 |
|
|
|
args.num_train_epochs = math.ceil(args.max_train_steps / num_update_steps_per_epoch) |
|
|
|
|
|
|
|
if accelerator.is_main_process: |
|
accelerator.init_trackers("dreambooth-lora-sd-xl", config=vars(args)) |
|
|
|
|
|
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 batches each epoch = {len(train_dataloader)}") |
|
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 |
|
|
|
|
|
if args.resume_from_checkpoint: |
|
if args.resume_from_checkpoint != "latest": |
|
path = os.path.basename(args.resume_from_checkpoint) |
|
else: |
|
|
|
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) |
|
|
|
|
|
progress_bar = tqdm(range(global_step, args.max_train_steps), disable=not accelerator.is_local_main_process) |
|
progress_bar.set_description("Steps") |
|
|
|
for epoch in range(first_epoch, args.num_train_epochs): |
|
unet.train() |
|
if args.train_text_encoder: |
|
text_encoder_one.train() |
|
text_encoder_two.train() |
|
for step, batch in enumerate(train_dataloader): |
|
|
|
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): |
|
pixel_values = batch["pixel_values"].to(dtype=vae.dtype) |
|
|
|
|
|
model_input = vae.encode(pixel_values).latent_dist.sample() |
|
model_input = model_input * vae.config.scaling_factor |
|
if args.pretrained_vae_model_name_or_path is None: |
|
model_input = model_input.to(weight_dtype) |
|
|
|
|
|
noise = torch.randn_like(model_input) |
|
bsz = model_input.shape[0] |
|
|
|
timesteps = torch.randint( |
|
0, noise_scheduler.config.num_train_timesteps, (bsz,), device=model_input.device |
|
) |
|
timesteps = timesteps.long() |
|
|
|
|
|
|
|
noisy_model_input = noise_scheduler.add_noise(model_input, noise, timesteps) |
|
|
|
|
|
elems_to_repeat = bsz // 2 if args.with_prior_preservation else bsz |
|
|
|
|
|
if not args.train_text_encoder: |
|
unet_added_conditions = { |
|
"time_ids": add_time_ids.repeat(elems_to_repeat, 1), |
|
"text_embeds": unet_add_text_embeds.repeat(elems_to_repeat, 1), |
|
} |
|
prompt_embeds_input = prompt_embeds.repeat(elems_to_repeat, 1, 1) |
|
model_pred = unet( |
|
noisy_model_input, |
|
timesteps, |
|
prompt_embeds_input, |
|
added_cond_kwargs=unet_added_conditions, |
|
).sample |
|
else: |
|
unet_added_conditions = {"time_ids": add_time_ids.repeat(elems_to_repeat, 1)} |
|
prompt_embeds, pooled_prompt_embeds = encode_prompt( |
|
text_encoders=[text_encoder_one, text_encoder_two], |
|
tokenizers=None, |
|
prompt=None, |
|
text_input_ids_list=[tokens_one, tokens_two], |
|
) |
|
unet_added_conditions.update({"text_embeds": pooled_prompt_embeds.repeat(elems_to_repeat, 1)}) |
|
prompt_embeds_input = prompt_embeds.repeat(elems_to_repeat, 1, 1) |
|
model_pred = unet( |
|
noisy_model_input, timesteps, prompt_embeds_input, added_cond_kwargs=unet_added_conditions |
|
).sample |
|
|
|
|
|
if noise_scheduler.config.prediction_type == "epsilon": |
|
target = noise |
|
elif noise_scheduler.config.prediction_type == "v_prediction": |
|
target = noise_scheduler.get_velocity(model_input, noise, timesteps) |
|
else: |
|
raise ValueError(f"Unknown prediction type {noise_scheduler.config.prediction_type}") |
|
|
|
if args.with_prior_preservation: |
|
|
|
model_pred, model_pred_prior = torch.chunk(model_pred, 2, dim=0) |
|
target, target_prior = torch.chunk(target, 2, dim=0) |
|
|
|
|
|
loss = F.mse_loss(model_pred.float(), target.float(), reduction="mean") |
|
|
|
|
|
prior_loss = F.mse_loss(model_pred_prior.float(), target_prior.float(), reduction="mean") |
|
|
|
|
|
loss = loss + args.prior_loss_weight * prior_loss |
|
else: |
|
loss = F.mse_loss(model_pred.float(), target.float(), reduction="mean") |
|
|
|
accelerator.backward(loss) |
|
if accelerator.sync_gradients: |
|
params_to_clip = ( |
|
itertools.chain(unet_lora_parameters, text_lora_parameters_one, text_lora_parameters_two) |
|
if args.train_text_encoder |
|
else unet_lora_parameters |
|
) |
|
accelerator.clip_grad_norm_(params_to_clip, args.max_grad_norm) |
|
optimizer.step() |
|
lr_scheduler.step() |
|
optimizer.zero_grad() |
|
|
|
|
|
if accelerator.sync_gradients: |
|
progress_bar.update(1) |
|
global_step += 1 |
|
|
|
if accelerator.is_main_process: |
|
if global_step % args.checkpointing_steps == 0: |
|
|
|
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])) |
|
|
|
|
|
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}") |
|
|
|
logs = {"loss": loss.detach().item(), "lr": lr_scheduler.get_last_lr()[0]} |
|
progress_bar.set_postfix(**logs) |
|
accelerator.log(logs, step=global_step) |
|
|
|
if global_step >= args.max_train_steps: |
|
break |
|
|
|
if accelerator.is_main_process: |
|
if args.validation_prompt is not None and epoch % args.validation_epochs == 0: |
|
logger.info( |
|
f"Running validation... \n Generating {args.num_validation_images} images with prompt:" |
|
f" {args.validation_prompt}." |
|
) |
|
|
|
if not args.train_text_encoder: |
|
text_encoder_one = text_encoder_cls_one.from_pretrained( |
|
args.pretrained_model_name_or_path, subfolder="text_encoder", revision=args.revision |
|
) |
|
text_encoder_two = text_encoder_cls_two.from_pretrained( |
|
args.pretrained_model_name_or_path, subfolder="text_encoder_2", revision=args.revision |
|
) |
|
pipeline = StableDiffusionXLPipeline.from_pretrained( |
|
args.pretrained_model_name_or_path, |
|
vae=vae, |
|
text_encoder=accelerator.unwrap_model(text_encoder_one), |
|
text_encoder_2=accelerator.unwrap_model(text_encoder_two), |
|
unet=accelerator.unwrap_model(unet), |
|
revision=args.revision, |
|
torch_dtype=weight_dtype, |
|
) |
|
|
|
|
|
scheduler_args = {} |
|
|
|
if "variance_type" in pipeline.scheduler.config: |
|
variance_type = pipeline.scheduler.config.variance_type |
|
|
|
if variance_type in ["learned", "learned_range"]: |
|
variance_type = "fixed_small" |
|
|
|
scheduler_args["variance_type"] = variance_type |
|
|
|
pipeline.scheduler = DPMSolverMultistepScheduler.from_config( |
|
pipeline.scheduler.config, **scheduler_args |
|
) |
|
|
|
pipeline = pipeline.to(accelerator.device) |
|
pipeline.set_progress_bar_config(disable=True) |
|
|
|
|
|
generator = torch.Generator(device=accelerator.device).manual_seed(args.seed) if args.seed else None |
|
pipeline_args = {"prompt": args.validation_prompt} |
|
|
|
with torch.cuda.amp.autocast(): |
|
images = [ |
|
pipeline(**pipeline_args, generator=generator).images[0] |
|
for _ in range(args.num_validation_images) |
|
] |
|
|
|
for tracker in accelerator.trackers: |
|
if tracker.name == "tensorboard": |
|
np_images = np.stack([np.asarray(img) for img in images]) |
|
tracker.writer.add_images("validation", np_images, epoch, dataformats="NHWC") |
|
if tracker.name == "wandb": |
|
tracker.log( |
|
{ |
|
"validation": [ |
|
wandb.Image(image, caption=f"{i}: {args.validation_prompt}") |
|
for i, image in enumerate(images) |
|
] |
|
} |
|
) |
|
|
|
del pipeline |
|
torch.cuda.empty_cache() |
|
|
|
|
|
accelerator.wait_for_everyone() |
|
if accelerator.is_main_process: |
|
unet = accelerator.unwrap_model(unet) |
|
unet = unet.to(torch.float32) |
|
unet_lora_layers = unet_attn_processors_state_dict(unet) |
|
|
|
if args.train_text_encoder: |
|
text_encoder_one = accelerator.unwrap_model(text_encoder_one) |
|
text_encoder_lora_layers = text_encoder_lora_state_dict(text_encoder_one.to(torch.float32)) |
|
text_encoder_two = accelerator.unwrap_model(text_encoder_two) |
|
text_encoder_2_lora_layers = text_encoder_lora_state_dict(text_encoder_two.to(torch.float32)) |
|
else: |
|
text_encoder_lora_layers = None |
|
text_encoder_2_lora_layers = None |
|
|
|
StableDiffusionXLPipeline.save_lora_weights( |
|
save_directory=args.output_dir, |
|
unet_lora_layers=unet_lora_layers, |
|
text_encoder_lora_layers=text_encoder_lora_layers, |
|
text_encoder_2_lora_layers=text_encoder_2_lora_layers, |
|
) |
|
|
|
|
|
|
|
vae = AutoencoderKL.from_pretrained( |
|
vae_path, |
|
subfolder="vae" if args.pretrained_vae_model_name_or_path is None else None, |
|
revision=args.revision, |
|
torch_dtype=weight_dtype, |
|
) |
|
pipeline = StableDiffusionXLPipeline.from_pretrained( |
|
args.pretrained_model_name_or_path, vae=vae, revision=args.revision, torch_dtype=weight_dtype |
|
) |
|
|
|
|
|
scheduler_args = {} |
|
|
|
if "variance_type" in pipeline.scheduler.config: |
|
variance_type = pipeline.scheduler.config.variance_type |
|
|
|
if variance_type in ["learned", "learned_range"]: |
|
variance_type = "fixed_small" |
|
|
|
scheduler_args["variance_type"] = variance_type |
|
|
|
pipeline.scheduler = DPMSolverMultistepScheduler.from_config(pipeline.scheduler.config, **scheduler_args) |
|
|
|
|
|
pipeline.load_lora_weights(args.output_dir) |
|
|
|
|
|
images = [] |
|
if args.validation_prompt and args.num_validation_images > 0: |
|
pipeline = pipeline.to(accelerator.device) |
|
generator = torch.Generator(device=accelerator.device).manual_seed(args.seed) if args.seed else None |
|
images = [ |
|
pipeline(args.validation_prompt, num_inference_steps=25, generator=generator).images[0] |
|
for _ in range(args.num_validation_images) |
|
] |
|
|
|
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) |
|
] |
|
} |
|
) |
|
|
|
if args.push_to_hub: |
|
save_model_card( |
|
repo_id, |
|
images=images, |
|
base_model=args.pretrained_model_name_or_path, |
|
train_text_encoder=args.train_text_encoder, |
|
prompt=args.instance_prompt, |
|
repo_folder=args.output_dir, |
|
vae_path=args.pretrained_vae_model_name_or_path, |
|
token=args.hub_token |
|
) |
|
upload_folder( |
|
repo_id=repo_id, |
|
folder_path=args.output_dir, |
|
commit_message="End of training", |
|
ignore_patterns=["step_*", "epoch_*"], |
|
token=args.hub_token |
|
) |
|
|
|
accelerator.end_training() |
|
|
|
|
|
if __name__ == "__main__": |
|
args = parse_args() |
|
main(args) |