VictorMylle/Kohya-ss-sd-scripts
1
0
Fork 1
mirror of https://github.com/kohya-ss/sd-scripts.git synced 2026-04-06 13:47:06 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix do not mean in batch dim when min_snr_gamma

Browse Source
This commit is contained in:
Kohya S
2023-03-27 20:47:11 +09:00
parent 14891523ce
commit 066b1bb57e
1 changed files with 16 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
fine_tune.py
View File

@@ -23,6 +23,7 @@ from library.config_util import (
import library.custom_train_functions as custom_train_functions
from library.custom_train_functions import apply_snr_weight
def train(args):
train_util.verify_training_args(args)
train_util.prepare_dataset_args(args, True)
@@ -62,8 +63,8 @@ def train(args):
blueprint = blueprint_generator.generate(user_config, args, tokenizer=tokenizer)
train_dataset_group = config_util.generate_dataset_group_by_blueprint(blueprint.dataset_group)
current_epoch = Value('i',0)
current_step = Value('i',0)
current_epoch = Value("i", 0)
current_step = Value("i", 0)
collater = train_util.collater_class(current_epoch, current_step)
if args.debug_dataset:
@@ -196,7 +197,9 @@ def train(args):
# 学習ステップ数を計算する
if args.max_train_epochs is not None:
args.max_train_steps = args.max_train_epochs * math.ceil(len(train_dataloader) / accelerator.num_processes / args.gradient_accumulation_steps)
args.max_train_steps = args.max_train_epochs * math.ceil(
len(train_dataloader) / accelerator.num_processes / args.gradient_accumulation_steps
)
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}")
# データセット側にも学習ステップを送信
@@ -308,10 +311,14 @@ def train(args):
else:
target = noise
loss = torch.nn.functional.mse_loss(noise_pred.float(), target.float(), reduction="mean")
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")
accelerator.backward(loss)
if accelerator.sync_gradients and args.max_grad_norm != 0.0:
@@ -407,7 +414,6 @@ def setup_parser() -> argparse.ArgumentParser:
config_util.add_config_arguments(parser)
custom_train_functions.add_custom_train_arguments(parser)
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")