Commit Graph

  • 0138a917d8 Update README.md Kohya S 2023-03-28 08:43:41 +09:00
  • 49b29f2db2 Merge pull request #333 from kohya-ss/dev Kohya S 2023-03-27 22:44:13 +09:00
  • 99eaf1fd65 fix typo Kohya S 2023-03-27 21:38:01 +09:00
  • 5fa20b5348 update readme Kohya S 2023-03-27 21:37:10 +09:00
  • 895b0b6ca7 Fix saving issue if epoch/step not in checkpoint Kohya S 2023-03-27 21:22:32 +09:00
  • 238f01bc9c fix images are used twice, update debug dataset Kohya S 2023-03-27 20:48:21 +09:00
  • 43a08b4061 add ja comment Kohya S 2023-03-27 20:47:27 +09:00
  • 066b1bb57e fix do not mean in batch dim when min_snr_gamma Kohya S 2023-03-27 20:47:11 +09:00
  • 3cdae0cbd2 Reduce peak RAM usage guaneec 2023-03-27 14:34:17 +08:00
  • 14891523ce fix seed for each dataset to make shuffling same Kohya S 2023-03-26 22:17:03 +09:00
  • 559a1aeeda Merge pull request #328 from mgz-dev/resize_lora-fixes Kohya S 2023-03-26 17:19:09 +09:00
  • a18558ddfe Merge pull request #308 from AI-Casanova/min-SNR Kohya S 2023-03-26 17:12:03 +09:00
  • 6732df93e2 Merge branch 'dev' into min-SNR Kohya S 2023-03-26 17:10:53 +09:00
  • 4f42f759ea Merge pull request #322 from u-haru/feature/token_warmup Kohya S 2023-03-26 17:05:59 +09:00
  • c9b157b536 update resize_lora.py (fix out of bounds and index) mgz-dev 2023-03-25 19:56:14 -05:00
  • 4c06bfad60 Fix for TypeError from bf16 precision: Thanks to mgz-dev AI-Casanova 2023-03-26 00:01:29 +00:00
  • a35d7ef227 Implement XTI Jakaline-dev 2023-03-26 05:26:10 +09:00
  • a4b34a9c3c blueprint_args_conflictは不要なため削除、shuffleが毎回行われる不具合修正 u-haru 2023-03-26 03:26:55 +09:00
  • 5a3d564a30 print削除 u-haru 2023-03-26 02:26:08 +09:00
  • 4dc1124f93 lora以外も対応 u-haru 2023-03-26 02:19:55 +09:00
  • 9c80da6ac5 Merge branch 'feature/token_warmup' of https://github.com/u-haru/sd-scripts into feature/token_warmup u-haru 2023-03-26 01:45:15 +09:00
  • 292cdb8379 データセットにepoch、stepが通達されないバグ修正 u-haru 2023-03-26 01:44:25 +09:00
  • 5ec90990de データセットにepoch、stepが通達されないバグ修正 u-haru 2023-03-26 01:41:24 +09:00
  • e203270e31 support TI embeds trained by WebUI(?) Kohya S 2023-03-24 20:46:42 +09:00
  • b2c5b96f2a format by black Kohya S 2023-03-24 20:19:05 +09:00
  • 1b89b2a10e シャッフル前にタグを切り詰めるように変更 u-haru 2023-03-24 13:44:30 +09:00
  • 143c26e552 競合時にpersistant_data_loader側を無効にするように変更 u-haru 2023-03-24 13:08:56 +09:00
  • 518a18aeff (ACTUAL) Min-SNR Weighting Strategy: Fixed SNR calculation to authors implementation AI-Casanova 2023-03-23 12:34:49 +00:00
  • a3c7d711e4 Min-SNR Weighting Strategy: Fixed SNR calculation to authors implementation AI-Casanova 2023-03-21 20:38:27 -05:00
  • dbadc40ec2 persistent_workersを有効にした際にキャプションが変化しなくなるバグ修正 u-haru 2023-03-23 12:33:03 +09:00
  • 447c56bf50 typo修正、stepをglobal_stepに修正、バグ修正 u-haru 2023-03-23 09:53:14 +09:00
  • a9b26b73e0 implement token warmup u-haru 2023-03-23 07:37:14 +09:00
  • 64c923230e Min-SNR Weighting Strategy: Refactored and added to all trainers AI-Casanova 2023-03-22 01:25:49 +00:00
  • 795a6bd2d8 Merge branch 'kohya-ss:main' into min-SNR AI-Casanova 2023-03-21 13:19:15 -05:00
  • aee343a9ee Merge pull request #310 from kohya-ss/dev Kohya S 2023-03-21 22:19:26 +09:00
  • 2c5949c155 update readme Kohya S 2023-03-21 22:17:20 +09:00
  • 193674e16c fix to support dynamic rank/alpha Kohya S 2023-03-21 21:59:51 +09:00
  • 4f92b6266c fix do not starting script Kohya S 2023-03-21 21:29:10 +09:00
  • 2d86f63e15 update steps calc with max_train_epochs Kohya S 2023-03-21 21:21:12 +09:00
  • 88751f58f6 Merge branch 'dev' of https://github.com/kohya-ss/sd-scripts into dev Kohya S 2023-03-21 21:10:44 +09:00
  • 7b324bcc3b support extensions of image files with uppercases Kohya S 2023-03-21 21:10:34 +09:00
  • 1645698ec0 Merge pull request #306 from robertsmieja/main Kohya S 2023-03-21 21:09:23 +09:00
  • 5aa5a07260 Merge pull request #292 from tsukimiya/hotfix/max_train_steps Kohya S 2023-03-21 21:02:29 +09:00
  • 6d9f3bc0b2 fix different reso in batch Kohya S 2023-03-21 18:33:46 +09:00
  • 1816ac3271 add vae_batch_size option for faster caching Kohya S 2023-03-21 18:15:57 +09:00
  • cca3804503 Merge branch 'main' into dev Kohya S 2023-03-21 15:05:41 +09:00
  • cb08fa0379 fix no npz with full path Kohya S 2023-03-21 15:05:25 +09:00
  • a265225972 Min-SNR Weighting Strategy AI-Casanova 2023-03-20 22:51:38 +00:00
  • eb66e5ebac Extract parser setup to helper function - Allows users who import the scripts to examine the parser programmatically Robert Smieja 2023-03-20 00:06:47 -04:00
  • 9d4cf8b03b Merge remote-tracking branch 'origin/hotfix/max_train_steps' into hotfix/max_train_steps tsukimiya 2023-03-19 23:55:51 +09:00
  • a167a592e2 Fixed an issue where max_train_steps was not set correctly when max_train_epochs was specified and gradient_accumulation_steps was set to 2 or more. tsukimiya 2023-03-13 14:37:28 +09:00
  • 432353185c Update README.md v0.5.2 Kohya S 2023-03-19 22:36:46 +09:00
  • d526f1d3d3 Merge pull request #305 from kohya-ss/dev Kohya S 2023-03-19 22:34:15 +09:00
  • c219600ca0 update readme Kohya S 2023-03-19 22:32:14 +09:00
  • de95431895 support win with diffusers, fix extra args eval Kohya S 2023-03-19 22:09:36 +09:00
  • c86bf213d1 Merge pull request #290 from orenwang/main Kohya S 2023-03-19 21:59:57 +09:00
  • 48c1be34f3 Merge branch 'dev' into main Kohya S 2023-03-19 21:58:41 +09:00
  • 140b4fad43 remove default values from output config Kohya S 2023-03-19 20:06:31 +09:00
  • 1f7babd2c7 Fix lpwp to support sdv2 and clip skip Kohya S 2023-03-19 11:10:17 +09:00
  • cfb19ad0da Merge pull request #288 from mio2333/main Kohya S 2023-03-19 10:57:47 +09:00
  • 1214760cea Merge branch 'dev' into main Kohya S 2023-03-19 10:56:56 +09:00
  • 64d85b2f51 fix num_processes, fix indent Kohya S 2023-03-19 10:52:46 +09:00
  • 8f08feb577 Merge pull request #271 from Isotr0py/dev Kohya S 2023-03-19 10:26:34 +09:00
  • ec7f9bab6c Merge branch 'dev' into dev Kohya S 2023-03-19 10:25:22 +09:00
  • 83e102c691 refactor config parse, feature to output config Kohya S 2023-03-19 10:11:11 +09:00
  • c3f9eb10f1 format with black Kohya S 2023-03-18 18:58:12 +09:00
  • 563a4dc897 Merge pull request #241 from Linaqruf/main Kohya S 2023-03-18 18:50:42 +09:00
  • 370ca9e8cd fix exception on training model in diffusers format orenwang 2023-03-13 14:32:22 +08:00
  • 5dad64b684 Fixed an issue where max_train_steps was not set correctly when max_train_epochs was specified and gradient_accumulation_steps was set to 2 or more. tsukimiya 2023-03-13 14:37:28 +09:00
  • e24a43ae0b sample images with weight and no length limit mio 2023-03-12 16:08:31 +08:00
  • 44d4cfb453 feat: added function to load training config with .toml Linaqruf 2023-03-12 11:52:37 +07:00
  • 7c1cf7f4ea Merge pull request #283 from kohya-ss/dev Kohya S 2023-03-11 08:05:30 +09:00
  • 0b38e663fd remove unnecessary device change Kohya S 2023-03-11 08:04:28 +09:00
  • 8b25929765 fix device error Kohya S 2023-03-11 08:03:02 +09:00
  • b80431de30 Merge pull request #278 from kohya-ss/dev v0.5.1 Kohya S 2023-03-10 22:05:36 +09:00
  • b177460807 restore comment Kohya S 2023-03-10 22:02:17 +09:00
  • c78c51c78f update documents Kohya S 2023-03-10 21:59:25 +09:00
  • 2652c9a66c Merge pull request #276 from mio2333/main Kohya S 2023-03-10 21:43:32 +09:00
  • 618592c52b npz check to use subset, add dadap warn close #274 Kohya S 2023-03-10 21:31:59 +09:00
  • 75d1883da6 fix LoRA rank is limited to target dim Kohya S 2023-03-10 21:12:15 +09:00
  • 4ad8e75291 fix to work with dim>320 Kohya S 2023-03-10 21:10:22 +09:00
  • e355b5e1d3 Merge pull request #269 from rvhfxb/patch-2 Kohya S 2023-03-10 20:56:11 +09:00
  • e3b2bb5b80 Merge branch 'dev' into dev Isotr0py 2023-03-10 19:04:07 +08:00
  • 7544b38635 fix multi gpu Isotr0py 2023-03-10 18:45:53 +08:00
  • 68cd874bb6 Append sys path for import_module mio 2023-03-10 18:29:34 +08:00
  • c4a596df9e replace unsafe eval() with ast Isotr0py 2023-03-10 13:44:16 +08:00
  • 00a9d734d9 Merge pull request #247 from ddPn08/dev Kohya S 2023-03-10 13:01:52 +09:00
  • 458173da5e Merge branch 'dev' into dev Kohya S 2023-03-10 13:00:49 +09:00
  • 1932c31c66 Merge pull request #243 from mgz-dev/dynamic-dim-lora-resize Kohya S 2023-03-10 12:59:39 +09:00
  • dd05d99efd Merge pull request #272 from kohya-ss/dev Kohya S 2023-03-09 21:54:41 +09:00
  • cf2bc437ec update readme Kohya S 2023-03-09 21:51:22 +09:00
  • aa317d4f57 Merge branch 'main' into dev Kohya S 2023-03-09 20:56:54 +09:00
  • 51249b1ba0 support conv2d 3x3 LoRA Kohya S 2023-03-09 20:56:33 +09:00
  • ab05be11d2 fix wrong typing Isotr0py 2023-03-09 19:35:06 +08:00
  • e7051d427c fix default conv alpha to 1 Kohya S 2023-03-09 20:26:14 +09:00
  • b885c6f9d2 disable annoying warning in CLIP loading Kohya S 2023-03-09 20:25:21 +09:00
  • ad443e172a fix samle gen failed if use templates Kohya S 2023-03-09 20:24:53 +09:00
  • eb68892ab1 add lr_scheduler_type etc Isotr0py 2023-03-09 16:51:22 +08:00
  • c4b4d1cb40 fix LoRA always expanded to Conv2d-3x3 Kohya S 2023-03-09 08:47:13 +09:00
  • 82aac26469 Update train_util.py rvhfxb 2023-03-08 22:42:41 +09:00