KohyaSS/fine_tune.py

431 lines
18 KiB
Python
Raw Normal View History

# training with captions
# XXX dropped option: hypernetwork training
import argparse
import gc
import math
import os
2023-03-20 12:47:00 +00:00
import toml
2023-03-28 15:54:42 +00:00
from multiprocessing import Value
from tqdm import tqdm
import torch
from accelerate.utils import set_seed
import diffusers
from diffusers import DDPMScheduler
import library.train_util as train_util
import library.config_util as config_util
from library.config_util import (
2023-03-20 12:47:00 +00:00
ConfigSanitizer,
BlueprintGenerator,
)
2023-03-28 15:54:42 +00:00
import library.custom_train_functions as custom_train_functions
from library.custom_train_functions import apply_snr_weight
def train(args):
2023-03-20 12:47:00 +00:00
train_util.verify_training_args(args)
train_util.prepare_dataset_args(args, True)
cache_latents = args.cache_latents
if args.seed is not None:
set_seed(args.seed) # 乱数系列を初期化する
tokenizer = train_util.load_tokenizer(args)
blueprint_generator = BlueprintGenerator(ConfigSanitizer(False, True, True))
if args.dataset_config is not None:
print(f"Load dataset config from {args.dataset_config}")
user_config = config_util.load_user_config(args.dataset_config)
ignored = ["train_data_dir", "in_json"]
if any(getattr(args, attr) is not None for attr in ignored):
print(
"ignore following options because config file is found: {0} / 設定ファイルが利用されるため以下のオプションは無視されます: {0}".format(
", ".join(ignored)
)
)
else:
2023-03-20 12:47:00 +00:00
user_config = {
"datasets": [
{
"subsets": [
{
"image_dir": args.train_data_dir,
"metadata_file": args.in_json,
}
]
}
]
}
blueprint = blueprint_generator.generate(user_config, args, tokenizer=tokenizer)
train_dataset_group = config_util.generate_dataset_group_by_blueprint(blueprint.dataset_group)
2023-03-28 15:54:42 +00:00
current_epoch = Value("i", 0)
current_step = Value("i", 0)
ds_for_collater = train_dataset_group if args.max_data_loader_n_workers == 0 else None
collater = train_util.collater_class(current_epoch, current_step, ds_for_collater)
2023-03-20 12:47:00 +00:00
if args.debug_dataset:
train_util.debug_dataset(train_dataset_group)
return
if len(train_dataset_group) == 0:
print(
"No data found. Please verify the metadata file and train_data_dir option. / 画像がありません。メタデータおよびtrain_data_dirオプションを確認してください。"
)
return
if cache_latents:
assert (
train_dataset_group.is_latent_cacheable()
), "when caching latents, either color_aug or random_crop cannot be used / latentをキャッシュするときはcolor_augとrandom_cropは使えません"
# acceleratorを準備する
print("prepare accelerator")
accelerator, unwrap_model = train_util.prepare_accelerator(args)
# mixed precisionに対応した型を用意しておき適宜castする
weight_dtype, save_dtype = train_util.prepare_dtype(args)
# モデルを読み込む
text_encoder, vae, unet, load_stable_diffusion_format = train_util.load_target_model(args, weight_dtype)
# verify load/save model formats
if load_stable_diffusion_format:
src_stable_diffusion_ckpt = args.pretrained_model_name_or_path
src_diffusers_model_path = None
else:
src_stable_diffusion_ckpt = None
src_diffusers_model_path = args.pretrained_model_name_or_path
2023-01-01 18:10:32 +00:00
2023-03-20 12:47:00 +00:00
if args.save_model_as is None:
save_stable_diffusion_format = load_stable_diffusion_format
use_safetensors = args.use_safetensors
else:
save_stable_diffusion_format = args.save_model_as.lower() == "ckpt" or args.save_model_as.lower() == "safetensors"
use_safetensors = args.use_safetensors or ("safetensors" in args.save_model_as.lower())
# Diffusers版のxformers使用フラグを設定する関数
def set_diffusers_xformers_flag(model, valid):
# model.set_use_memory_efficient_attention_xformers(valid) # 次のリリースでなくなりそう
# pipeが自動で再帰的にset_use_memory_efficient_attention_xformersを探すんだって(;´Д`)
# U-Netだけ使う時にはどうすればいいのか……仕方ないからコピって使うか
# 0.10.2でなんか巻き戻って個別に指定するようになった(;^ω^)
# Recursively walk through all the children.
# Any children which exposes the set_use_memory_efficient_attention_xformers method
# gets the message
def fn_recursive_set_mem_eff(module: torch.nn.Module):
if hasattr(module, "set_use_memory_efficient_attention_xformers"):
module.set_use_memory_efficient_attention_xformers(valid)
for child in module.children():
fn_recursive_set_mem_eff(child)
fn_recursive_set_mem_eff(model)
# モデルに xformers とか memory efficient attention を組み込む
if args.diffusers_xformers:
print("Use xformers by Diffusers")
set_diffusers_xformers_flag(unet, True)
else:
# Windows版のxformersはfloatで学習できないのでxformersを使わない設定も可能にしておく必要がある
print("Disable Diffusers' xformers")
set_diffusers_xformers_flag(unet, False)
train_util.replace_unet_modules(unet, args.mem_eff_attn, args.xformers)
# 学習を準備する
if cache_latents:
vae.to(accelerator.device, dtype=weight_dtype)
vae.requires_grad_(False)
vae.eval()
with torch.no_grad():
2023-03-22 00:20:57 +00:00
train_dataset_group.cache_latents(vae, args.vae_batch_size)
2023-03-20 12:47:00 +00:00
vae.to("cpu")
if torch.cuda.is_available():
torch.cuda.empty_cache()
gc.collect()
# 学習を準備する:モデルを適切な状態にする
training_models = []
if args.gradient_checkpointing:
unet.enable_gradient_checkpointing()
training_models.append(unet)
if args.train_text_encoder:
print("enable text encoder training")
if args.gradient_checkpointing:
text_encoder.gradient_checkpointing_enable()
training_models.append(text_encoder)
else:
text_encoder.to(accelerator.device, dtype=weight_dtype)
text_encoder.requires_grad_(False) # text encoderは学習しない
if args.gradient_checkpointing:
text_encoder.gradient_checkpointing_enable()
text_encoder.train() # required for gradient_checkpointing
2023-01-01 18:10:32 +00:00
else:
2023-03-20 12:47:00 +00:00
text_encoder.eval()
2023-03-20 12:47:00 +00:00
if not cache_latents:
vae.requires_grad_(False)
vae.eval()
vae.to(accelerator.device, dtype=weight_dtype)
2023-03-20 12:47:00 +00:00
for m in training_models:
m.requires_grad_(True)
params = []
for m in training_models:
params.extend(m.parameters())
params_to_optimize = params
# 学習に必要なクラスを準備する
print("prepare optimizer, data loader etc.")
_, _, optimizer = train_util.get_optimizer(args, trainable_params=params_to_optimize)
# dataloaderを準備する
# DataLoaderのプロセス数0はメインプロセスになる
n_workers = min(args.max_data_loader_n_workers, os.cpu_count() - 1) # cpu_count-1 ただし最大で指定された数まで
train_dataloader = torch.utils.data.DataLoader(
train_dataset_group,
batch_size=1,
shuffle=True,
2023-03-28 15:54:42 +00:00
collate_fn=collater,
2023-03-20 12:47:00 +00:00
num_workers=n_workers,
persistent_workers=args.persistent_data_loader_workers,
)
# 学習ステップ数を計算する
if args.max_train_epochs is not None:
2023-03-28 15:54:42 +00:00
args.max_train_steps = args.max_train_epochs * math.ceil(
len(train_dataloader) / accelerator.num_processes / args.gradient_accumulation_steps
)
2023-03-20 12:47:00 +00:00
print(f"override steps. steps for {args.max_train_epochs} epochs is / 指定エポックまでのステップ数: {args.max_train_steps}")
2023-03-28 15:54:42 +00:00
# データセット側にも学習ステップを送信
train_dataset_group.set_max_train_steps(args.max_train_steps)
2023-03-20 12:47:00 +00:00
# lr schedulerを用意する
lr_scheduler = train_util.get_scheduler_fix(args, optimizer, accelerator.num_processes)
# 実験的機能勾配も含めたfp16学習を行う モデル全体をfp16にする
if args.full_fp16:
assert (
args.mixed_precision == "fp16"
), "full_fp16 requires mixed precision='fp16' / full_fp16を使う場合はmixed_precision='fp16'を指定してください。"
print("enable full fp16 training.")
unet.to(weight_dtype)
text_encoder.to(weight_dtype)
# acceleratorがなんかよろしくやってくれるらしい
if args.train_text_encoder:
unet, text_encoder, optimizer, train_dataloader, lr_scheduler = accelerator.prepare(
unet, text_encoder, optimizer, train_dataloader, lr_scheduler
)
else:
unet, optimizer, train_dataloader, lr_scheduler = accelerator.prepare(unet, optimizer, train_dataloader, lr_scheduler)
# 実験的機能勾配も含めたfp16学習を行う PyTorchにパッチを当ててfp16でのgrad scaleを有効にする
if args.full_fp16:
train_util.patch_accelerator_for_fp16_training(accelerator)
# resumeする
if args.resume is not None:
print(f"resume training from state: {args.resume}")
accelerator.load_state(args.resume)
# epoch数を計算する
num_update_steps_per_epoch = math.ceil(len(train_dataloader) / args.gradient_accumulation_steps)
num_train_epochs = math.ceil(args.max_train_steps / num_update_steps_per_epoch)
if (args.save_n_epoch_ratio is not None) and (args.save_n_epoch_ratio > 0):
args.save_every_n_epochs = math.floor(num_train_epochs / args.save_n_epoch_ratio) or 1
# 学習する
total_batch_size = args.train_batch_size * accelerator.num_processes * args.gradient_accumulation_steps
print("running training / 学習開始")
print(f" num examples / サンプル数: {train_dataset_group.num_train_images}")
print(f" num batches per epoch / 1epochのバッチ数: {len(train_dataloader)}")
print(f" num epochs / epoch数: {num_train_epochs}")
print(f" batch size per device / バッチサイズ: {args.train_batch_size}")
print(f" total train batch size (with parallel & distributed & accumulation) / 総バッチサイズ(並列学習、勾配合計含む): {total_batch_size}")
2023-03-22 00:20:57 +00:00
print(f" gradient accumulation steps / 勾配を合計するステップ数 = {args.gradient_accumulation_steps}")
2023-03-20 12:47:00 +00:00
print(f" total optimization steps / 学習ステップ数: {args.max_train_steps}")
progress_bar = tqdm(range(args.max_train_steps), smoothing=0, disable=not accelerator.is_local_main_process, desc="steps")
global_step = 0
noise_scheduler = DDPMScheduler(
beta_start=0.00085, beta_end=0.012, beta_schedule="scaled_linear", num_train_timesteps=1000, clip_sample=False
)
if accelerator.is_main_process:
accelerator.init_trackers("finetuning")
for epoch in range(num_train_epochs):
print(f"epoch {epoch+1}/{num_train_epochs}")
2023-03-28 15:54:42 +00:00
current_epoch.value = epoch + 1
2023-03-20 12:47:00 +00:00
for m in training_models:
m.train()
loss_total = 0
for step, batch in enumerate(train_dataloader):
2023-03-28 15:54:42 +00:00
current_step.value = global_step
2023-03-20 12:47:00 +00:00
with accelerator.accumulate(training_models[0]): # 複数モデルに対応していない模様だがとりあえずこうしておく
with torch.no_grad():
if "latents" in batch and batch["latents"] is not None:
latents = batch["latents"].to(accelerator.device)
else:
# latentに変換
latents = vae.encode(batch["images"].to(dtype=weight_dtype)).latent_dist.sample()
latents = latents * 0.18215
b_size = latents.shape[0]
with torch.set_grad_enabled(args.train_text_encoder):
# Get the text embedding for conditioning
input_ids = batch["input_ids"].to(accelerator.device)
encoder_hidden_states = train_util.get_hidden_states(
args, input_ids, tokenizer, text_encoder, None if not args.full_fp16 else weight_dtype
)
# Sample noise that we'll add to the latents
noise = torch.randn_like(latents, device=latents.device)
if args.noise_offset:
# https://www.crosslabs.org//blog/diffusion-with-offset-noise
noise += args.noise_offset * torch.randn((latents.shape[0], latents.shape[1], 1, 1), device=latents.device)
# Sample a random timestep for each image
timesteps = torch.randint(0, noise_scheduler.config.num_train_timesteps, (b_size,), device=latents.device)
timesteps = timesteps.long()
# Add noise to the latents according to the noise magnitude at each timestep
# (this is the forward diffusion process)
noisy_latents = noise_scheduler.add_noise(latents, noise, timesteps)
# Predict the noise residual
noise_pred = unet(noisy_latents, timesteps, encoder_hidden_states).sample
if args.v_parameterization:
# v-parameterization training
target = noise_scheduler.get_velocity(latents, noise, timesteps)
else:
target = noise
2023-03-28 15:54:42 +00:00
if args.min_snr_gamma:
# do not mean over batch dimension for snr weight
loss = torch.nn.functional.mse_loss(noise_pred.float(), target.float(), reduction="none")
loss = loss.mean([1, 2, 3])
loss = apply_snr_weight(loss, timesteps, noise_scheduler, args.min_snr_gamma)
loss = loss.mean() # mean over batch dimension
else:
loss = torch.nn.functional.mse_loss(noise_pred.float(), target.float(), reduction="mean")
2023-03-20 12:47:00 +00:00
accelerator.backward(loss)
if accelerator.sync_gradients and args.max_grad_norm != 0.0:
params_to_clip = []
for m in training_models:
params_to_clip.extend(m.parameters())
accelerator.clip_grad_norm_(params_to_clip, args.max_grad_norm)
optimizer.step()
lr_scheduler.step()
optimizer.zero_grad(set_to_none=True)
# Checks if the accelerator has performed an optimization step behind the scenes
if accelerator.sync_gradients:
progress_bar.update(1)
global_step += 1
train_util.sample_images(
accelerator, args, None, global_step, accelerator.device, vae, tokenizer, text_encoder, unet
)
current_loss = loss.detach().item() # 平均なのでbatch sizeは関係ないはず
if args.logging_dir is not None:
logs = {"loss": current_loss, "lr": float(lr_scheduler.get_last_lr()[0])}
if args.optimizer_type.lower() == "DAdaptation".lower(): # tracking d*lr value
logs["lr/d*lr"] = (
lr_scheduler.optimizers[0].param_groups[0]["d"] * lr_scheduler.optimizers[0].param_groups[0]["lr"]
)
accelerator.log(logs, step=global_step)
# TODO moving averageにする
loss_total += current_loss
avr_loss = loss_total / (step + 1)
logs = {"loss": avr_loss} # , "lr": lr_scheduler.get_last_lr()[0]}
progress_bar.set_postfix(**logs)
if global_step >= args.max_train_steps:
break
if args.logging_dir is not None:
logs = {"loss/epoch": loss_total / len(train_dataloader)}
accelerator.log(logs, step=epoch + 1)
accelerator.wait_for_everyone()
if args.save_every_n_epochs is not None:
src_path = src_stable_diffusion_ckpt if save_stable_diffusion_format else src_diffusers_model_path
train_util.save_sd_model_on_epoch_end(
args,
accelerator,
src_path,
save_stable_diffusion_format,
use_safetensors,
save_dtype,
epoch,
num_train_epochs,
global_step,
unwrap_model(text_encoder),
unwrap_model(unet),
vae,
)
train_util.sample_images(accelerator, args, epoch + 1, global_step, accelerator.device, vae, tokenizer, text_encoder, unet)
is_main_process = accelerator.is_main_process
if is_main_process:
unet = unwrap_model(unet)
text_encoder = unwrap_model(text_encoder)
accelerator.end_training()
if args.save_state:
train_util.save_state_on_train_end(args, accelerator)
del accelerator # この後メモリを使うのでこれは消す
if is_main_process:
src_path = src_stable_diffusion_ckpt if save_stable_diffusion_format else src_diffusers_model_path
train_util.save_sd_model_on_train_end(
args, src_path, save_stable_diffusion_format, use_safetensors, save_dtype, epoch, global_step, text_encoder, unet, vae
)
print("model saved.")
2023-03-22 00:20:57 +00:00
def setup_parser() -> argparse.ArgumentParser:
2023-03-20 12:47:00 +00:00
parser = argparse.ArgumentParser()
train_util.add_sd_models_arguments(parser)
train_util.add_dataset_arguments(parser, False, True, True)
train_util.add_training_arguments(parser, False)
train_util.add_sd_saving_arguments(parser)
train_util.add_optimizer_arguments(parser)
config_util.add_config_arguments(parser)
2023-03-28 15:54:42 +00:00
custom_train_functions.add_custom_train_arguments(parser)
2023-03-20 12:47:00 +00:00
parser.add_argument("--diffusers_xformers", action="store_true", help="use xformers by diffusers / Diffusersでxformersを使用する")
parser.add_argument("--train_text_encoder", action="store_true", help="train text encoder / text encoderも学習する")
2023-03-22 00:20:57 +00:00
return parser
if __name__ == "__main__":
parser = setup_parser()
2023-03-20 12:47:00 +00:00
args = parser.parse_args()
args = train_util.read_config_from_file(args, parser)
train(args)