Kohya S a8643095ca fix: LECO PR #2285 のレビュー指摘事項を修正 ...

- train_util.py/deepspeed_utils.py の getattr 化を元に戻し、LECO パーサーにダミー引数を追加
- sdxl_train_util のモジュールレベルインポートをローカルインポートに変更
- PromptEmbedsCache.__getitem__ でキャッシュミス時に KeyError を送出するよう修正
- 設定ファイル形式を YAML から TOML に変更（リポジトリの規約に統一）
- 重複コード (build_network_kwargs, get_save_extension, save_weights) を leco_train_util.py に統合
- _expand_slider_target の冗長な PromptSettings 構築を簡素化
- add_time_ids 用に専用の batch_add_time_ids 関数を追加

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

2026-03-28 20:06:34 +09:00

..

anima_train_network.md

Fix the LoRA dropout issue in the Anima model during LoRA training. (#2272)

2026-02-23 15:13:40 +09:00

config_README-en.md

docs: add skip_image_resolution to config README (#2288)

2026-03-19 09:17:29 +09:00

config_README-ja.md

docs: add skip_image_resolution to config README (#2288)

2026-03-19 09:17:29 +09:00

fine_tune_README_ja.md

move documents to docs folder

2023-05-07 20:56:42 +09:00

fine_tune.md

doc: add comprehensive fine-tuning guide for various model architectures

2025-09-01 12:36:51 +09:00

flux_train_network.md

Merge branch 'sd3' into doc-update-for-latest-features

2025-09-09 12:42:09 +09:00

gen_img_README-ja.md

feat: add pyramid noise and noise offset options to generation script (#2252)

2026-01-18 16:56:48 +09:00

gen_img_README.md

feat: add pyramid noise and noise offset options to generation script (#2252)

2026-01-18 16:56:48 +09:00

hunyuan_image_train_network.md

docs: enhance text encoder CPU usage instructions for HunyuanImage-2.1 training

2025-09-28 18:21:25 +09:00

loha_lokr.md

feat: Support LoKr/LoHa for SDXL and Anima (#2275)

2026-02-23 22:09:00 +09:00

lumina_train_network.md

doc: add Sage Attention and sample batch size options to Lumina training guide

2025-08-31 21:19:28 +09:00

masked_loss_README-ja.md

update docs for masked loss

2024-05-27 21:25:06 +09:00

masked_loss_README.md

update docs for masked loss

2024-05-27 21:25:06 +09:00

sd3_train_network.md

Doc update sd3 branch documentation (#2253)

2026-01-19 21:38:46 +09:00

sdxl_train_network.md

update docs. sdxl is transltaed, flux.1 is corrected

2025-04-13 22:06:58 +09:00

train_db_README-ja.md

move documents to docs folder

2023-05-07 20:56:42 +09:00

train_db_README-zh.md

move documents to docs folder

2023-05-07 20:56:42 +09:00

train_leco.md

fix: LECO PR #2285 のレビュー指摘事項を修正

2026-03-28 20:06:34 +09:00

train_lllite_README-ja.md

common masked loss func, apply to all training script

2024-03-17 19:30:20 +09:00

train_lllite_README.md

update docs and help text

2024-10-11 22:23:15 +09:00

train_network_advanced.md

Doc update sd3 branch documentation (#2253)

2026-01-19 21:38:46 +09:00

train_network_README-ja.md

画像のアルファチャンネルをlossのマスクとして使用するオプションを追加 (#1223)

2024-05-19 19:07:25 +09:00

train_network_README-zh.md

画像のアルファチャンネルをlossのマスクとして使用するオプションを追加 (#1223)

2024-05-19 19:07:25 +09:00

train_network.md

Doc update sd3 branch documentation (#2253)

2026-01-19 21:38:46 +09:00

train_README-ja.md

correct option name closes #1446

2024-08-11 21:54:10 +09:00

train_README-zh.md

correct option name closes #1446

2024-08-11 21:54:10 +09:00

train_SDXL-en.md

update PyTorch version and reorganize dependencies

2024-03-24 22:06:37 +09:00

train_textual_inversion.md

Update docs/train_textual_inversion.md

2025-09-04 19:46:04 +09:00

train_ti_README-ja.md

move documents to docs folder

2023-05-07 20:56:42 +09:00

validation.md

doc: add validation loss documentation for model training

2025-08-31 21:39:47 +09:00

wd14_tagger_README-en.md

fix: metadata dataset degradation and make it work (#2186)

2026-01-18 15:17:07 +09:00

wd14_tagger_README-ja.md

fix: metadata dataset degradation and make it work (#2186)

2026-01-18 15:17:07 +09:00