1580158
/
kohya_ss
mirror of https://github.com/bmaltais/kohya_ss
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
3,216 Commits
16 Branches
178 Tags
68 MiB
Commit Graph

303 Commits (8fb9001a2c866a6796c8a34f3e7bb9d2d99ddb6a)

Author SHA1 Message Date
Koro 781e7e5f48
fixed names on LR Schedure dropdown (#3064) 2025-03-28 08:55:51 -04:00
bmaltais 954a8c7fb4 Fix dreambooth issue 2025-02-24 19:27:26 -05:00
bmaltais 836417c3bb Fix issue with SD3 Lora training blocks_to_swap and fused_backward_pass 2025-02-24 18:36:13 -05:00
bmaltais 6299e82905 Add uniform to weighting_scheme 2024-12-31 09:57:52 -05:00
bmaltais a10986ad30 Fix network_module issue 2024-12-31 09:46:42 -05:00
bmaltais 8afcaf8be7 Fix train_t5xxl issue 2024-12-31 09:43:34 -05:00
bmaltais 4fd25ced85 Fix clip_l issue for missing path 2024-12-31 09:35:36 -05:00
bmaltais 65da590eab Add a bunch of missing SD3 parameters 2024-12-31 08:48:48 -05:00
bmaltais 4a741a8a4a Add support for sd3 lora disable_mmap_load_safetensors 2024-12-31 08:24:37 -05:00
bmaltais 3c860c4035 Make blocks_to_swap common 2024-12-31 08:11:42 -05:00
bmaltais 3eec4c9b32 SD3 LoRA training MVP 2024-12-31 07:59:55 -05:00
bmaltais 730ed13546 Add support for prodigyplus.ProdigyPlusScheduleFree 2024-12-31 07:12:43 -05:00
bmaltais 7068c7dabe Add support for fused_backward_pass for sd3 finetuning 2024-12-30 19:37:03 -05:00
bmaltais fce89ad798 Add support for huber_scale 2024-12-30 19:21:03 -05:00
bmaltais 45cfc1d868 Add support for RAdamScheduleFree 2024-12-30 19:13:20 -05:00
bmaltais d0cd9f547a Reverting all changes for max_train_steps 2024-11-18 19:31:51 -05:00
bmaltais 309a9bbc1b Fix max_train_steps_info error 2024-11-18 07:07:33 -05:00
bmaltais a6f0ff72a8 Fix issue with max_train_steps 2024-11-17 18:37:27 -05:00
bmaltais b332b1df9d Fix blocks_to_swap not properly working 2024-11-16 20:33:48 -05:00
bmaltais 0d8fe4db4c Make blocks_to_swap visible in LoRA tab 2024-11-16 08:24:16 -05:00
bmaltais a35d0cdccb Add support for quiet output for linux setup 2024-11-07 10:33:29 -05:00
bmaltais 021c6f5ae3 Add support for save_last_n_epochs 2024-10-29 20:14:02 -04:00
bmaltais b77da103c4 Fix typo 2024-10-26 19:47:25 -04:00
b-fission d29a546c9a Increase maximum flux-lora merge strength to 2 2024-10-21 16:38:53 -05:00
bmaltais b9fe58cd1e Add support for --save_last_n_epochs_state 2024-10-18 17:53:06 -04:00
bmaltais 5e10c380d4 Add support for LyCORIS LoRA when training Flux.1 2024-10-17 18:51:15 -04:00
bmaltais 3c3f1edfb6 Add --skip_cache_check option to GUI 2024-10-14 18:32:09 -04:00
bmaltais 55cd56a63b Allow to specify tensorboard host via env var TENSORBOARD_HOST 2024-10-09 21:45:31 -04:00
bmaltais faa3d1920c Update Lycoris support 2024-10-06 20:26:33 -04:00
bmaltais 34db05e8fd Add support for AdEMAMix8bit optimizer 2024-09-29 13:41:57 -04:00
bmaltais a28026bca1 Adding support for blocks_to_swap option to gui 2024-09-29 08:36:05 -04:00
bmaltais cee863664b Add support for ScheduleFree Optimizer Type 2024-09-17 19:57:41 -04:00
bmaltais 0ba8ab6c71
Merge pull request #2829 from iamrohitanshu/sd3-flux.1 
Update 3 files in kohya_gui folder
 2024-09-17 18:27:13 -04:00
bmaltais e696df27df Fix issue with old_lr_warmup_steps 2024-09-17 18:06:42 -04:00
bmaltais 8c11a08eb3 Add missing network parms 2024-09-16 18:51:09 -04:00
bmaltais 416ef0e19e Add support for Blocks to train 2024-09-16 18:48:07 -04:00
rohitanshu b8c4de17e5
Update textual_inversion_gui.py 
Fixed minor typos of "Regularization"
 2024-09-16 19:45:41 +05:30
rohitanshu f8ddd596d6
Update dreambooth_gui.py 
Fixed minor typos of "Regularization"
 2024-09-16 19:44:51 +05:30
rohitanshu 9a37a0e197
Update lora_gui.py 
Fixed minor typos of "Regularization"
 2024-09-16 19:44:05 +05:30
bmaltais d24fae17b7 Add support for Rank for layers 2024-09-14 13:45:01 -04:00
bmaltais c554984a53 Fix logic typo 2024-09-14 09:24:13 -04:00
bmaltais 31c77d6d71 Preliminary support for Flux1 OFT 2024-09-14 09:13:26 -04:00
bmaltais f744479f9a Revert some changes 2024-09-13 15:53:30 -04:00
bmaltais 46b57ed635 Fix for max_train_steps 2024-09-13 15:35:30 -04:00
bmaltais 0d53ac6676 Removing stable-diffusion-1.5 default model 2024-09-12 21:57:20 -04:00
bmaltais 20d7e659e1 Update lr_warmup_steps code 2024-09-12 21:31:18 -04:00
bmaltais e655ebf2dd Add support for lr_warmup_steps override 2024-09-12 20:56:05 -04:00
bmaltais 7c190edef8 Fix issue with new text_encoder_lr parameter syntax 2024-09-12 07:03:45 -04:00
bmaltais 3cc33f497f Update transformers and wandb module 2024-09-11 19:39:53 -04:00
bmaltais f365b630ff Update sd-scripts and add support for t5xxl LR 2024-09-11 19:17:37 -04:00