Pull Requests kohya-ss/sd-scripts

feat: SD1.x/2.x と SDXL 向けの LECO 学習スクリプトを追加

#2285 by umisetokikaze was merged 2026-03-28 10:25

Codex/improve training script pipeline qjq4be

#2284 by orangemagic123 was closed 2026-03-14 04:52

Bump `pytorch-optimizer` into 3.10.0

#2283 by kozistr was merged 2026-03-19 00:18

fix: add v-prediction support to DDPM

#2281 by Maru-mee was closed 2026-03-01 06:41

fix: rename character_tags to img_character_tags to fix UnboundLocalError

#2280 by kohya-ss was merged 2026-02-25 23:21

feat: Stability with fp16 for anima

#2277 by kohya-ss was merged 2026-02-23 12:15

feat: backward compatibility for SD/SDXL latent cache

#2276 by kohya-ss was merged 2026-02-23 12:44

feat: Support LoKr/LoHa for SDXL and Anima

#2275 by kohya-ss was merged 2026-02-23 13:09

feat: add fp16 safe patch option for training on older GPUs to prevent NaNs for Anima model

#2274 by ihatenumbers was closed 2026-02-22 14:35

Add `skip_image_resolution` to deduplicate multi-resolution dataset

#2273 by woct0rdho was merged 2026-03-18 23:43

Fix the LoRA dropout issue in the Anima model and implement support for network_reg_alphas during LoRA training.

#2272 by duongve13112002 was merged 2026-02-23 06:13

Deduplicate multi-resolution dataset with `bucket_no_upscale`

#2270 by woct0rdho was closed 2026-02-20 06:20

Multi-resolution dataset for SD1/SDXL

#2269 by woct0rdho was merged 2026-02-23 06:30

merge sd3 to main

#2268 by kohya-ss was merged 2026-02-15 23:07

fix: `str is not "no"` to `str != "no"`

#2267 by kohya-ss was merged 2026-02-15 23:01

Release 0.10.1

#2264 by kohya-ss was merged 2026-02-12 23:34

feat: Anima support

#2263 by kohya-ss was merged 2026-02-12 23:20

Fix bug and optimization for Lumina model

#2262 by duongve13112002 was merged 2026-02-15 22:54

Add/modify some implementation for anima

#2261 by kohya-ss was merged 2026-02-12 23:15

Support Anima model

#2260 by duongve13112002 was merged 2026-02-08 01:18

Fix AttributeError for UNet2DConditionModel with newer diffusers versions

#2255 by cgcalatrava was merged 2026-01-19 22:50

Merge the changes from the sd3 branch into main

#2254 by kohya-ss was merged 2026-01-19 13:00

Doc update sd3 branch documentation

#2253 by kohya-ss was merged 2026-01-19 12:38

feat: add pyramid noise and noise offset options to generation script

#2252 by kohya-ss was merged 2026-01-18 07:56

fix(tests): add ip_noise_gamma args for MockArgs in pytest

#2251 by kohya-ss was merged 2026-01-18 06:09

Bump `pytorch-optimizer` version to v3.9.0

#2246 by kozistr was merged 2025-12-21 13:51

Use `torch.svd_lowrank` for large matrices in `resize_lora.py`

#2240 by woct0rdho was merged 2026-03-29 10:54

Feature/te lr after

#2232 by maruo555 was closed 2025-11-03 05:52

fix: error on batch generation closes #2209

#2226 by kohya-ss was merged 2025-10-15 12:57

fix: lumina 2 timesteps handling

#2225 by urlesistiana was merged 2026-01-18 05:39