nodes/kijai-ComfyUI-FluxTrainer
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
121 Commits 4 Branches 0 Tags
54d8a4b60dfe88ead02fa17490bf8d96908a1fe4
Commit Graph
40 Commits
Author SHA1 Message Date
kijai 95de0cbcc9 allow loading models saved in Comfy 2024-09-08 15:57:45 +03:00
kijai 1c70cce963 support fp8 e5m2 2024-09-07 18:27:02 +03:00
kijai a4d2f7545b Update flux_train_utils.py 2024-09-04 19:42:30 +03:00
kijai 38d3d9ddf6 initial T5 training 2024-09-04 18:50:27 +03:00
kijai 16464ff1ba fix flux_shift 2024-09-01 18:46:45 +03:00
kijai 8f6a90e5a9 add flux_shift 2024-09-01 13:37:13 +03:00
kijai 816f0ec75f fix bug with dataset adding stacking 
removing a dataset node still won't remove added dataset
 2024-08-30 03:11:15 +03:00
kijai aba4a61024 Update train_util.py 2024-08-30 01:44:22 +03:00
kijai fc1f0e59e2 better error for invalid dataset path 2024-08-30 01:36:52 +03:00
kijai 66749c8772 clip training 2024-08-29 12:48:33 +03:00
kijai e03b35677e expose some more args, add shift -timestep sampling method 
https://github.com/kohya-ss/sd-scripts/commit/72287d39c76176c0e1c16e8da4f5ddc6f94ea7d6
 2024-08-25 14:18:50 +03:00
kijai 9f52607be1 class_token first instead 2024-08-24 18:48:04 +03:00
kijai 1e20da3eca Add LoRA extraction node 2024-08-24 11:47:34 +03:00
kijai 43d930c462 full fine tune fixes 2024-08-23 21:41:29 +03:00
kijai 1ba1ef0ec1 Update deepspeed_utils.py 2024-08-23 01:39:04 +03:00
kijai 74d6d35ca1 fixes from upstream 2024-08-22 20:42:19 +03:00
kijai 3829715515 Update train_util.py 2024-08-22 19:45:20 +03:00
kijai 957ac37c5d always add class_token if provided 2024-08-22 19:40:27 +03:00
Kijai 14b2d50e22 Check output folder before starting and save args as .json there. 2024-08-21 18:40:28 +03:00
Kijai 9444038bda finetune fixes from upstream 2024-08-21 17:21:15 +03:00
Kijai 617c33da0f attn mask bugfix 2024-08-21 16:08:46 +03:00
kijai 9f6f596c0d t5_attn_mask update from upstream 2024-08-21 04:26:25 +03:00
kijai 4ddfd9f70b fix more imports 2024-08-20 20:53:34 +03:00
kijai 9aea8e5137 fix imports 2024-08-20 20:05:24 +03:00
kijai 2011e0924a Update strategy_base.py 2024-08-20 19:59:26 +03:00
kijai a559358ac8 revert to kohya's multires training 2024-08-20 19:52:33 +03:00
kijai 65aa2df52c add CAME 2024-08-20 17:13:54 +03:00
kijai d07378a99d updates 2024-08-20 16:34:46 +03:00
Kijai 2a3dcae345 loss visualizer fixes 2024-08-19 17:54:20 +03:00
kijai 2cd3c8e170 fix loss visualizer, and other small stuff 2024-08-19 02:51:19 +03:00
kijai fdcb5a8686 pbar 2024-08-18 22:03:33 +03:00
kijai e271bbd8f1 updates 2024-08-18 21:40:14 +03:00
kijai c8be524730 updates 2024-08-18 02:10:18 +03:00
kijai 7678caf9c7 multiple max bucket resos 2024-08-17 15:59:32 +03:00
kijai 0fb8a0030a add inference node 2024-08-17 05:01:23 +03:00
kijai 120a986cf0 Update train_util.py 2024-08-16 19:15:28 +03:00
Kijai bf4b91a818 update 2024-08-16 18:17:01 +03:00
kijai dc20750105 fixes 2024-08-16 12:17:11 +03:00
Kijai 7f5ca3c809 updates 2024-08-15 17:38:31 +03:00
kijai 596149fc8d init 2024-08-15 11:32:11 +03:00