|
2eb809a0c8
|
Add lora cache mode in comfyui && Use cuda in merge_lora func for speeding up. (#172)
* fix lora precision problem during multiple merge unmerge, keep the original weights on cpu
* add comments
* update
* Update Lora cache
---------
Co-authored-by: mengli.cml <mengli.cml@alibaba-inc.com>
Co-authored-by: bubbliiiing <3323290568@qq.com>
|
2024-12-24 10:02:57 +08:00 |
|
|
hkzandbubbliiiing
|
17c1f02ba8
|
Add reward LoRA training (#160)
* Add reward lora
* Fix bug in lora utils && Update Readme
---------
Co-authored-by: bubbliiiing <3323290568@qq.com>
|
2024-11-28 13:59:53 +08:00 |
|
|
Bubbliiiing
|
62de94e2f1
|
Update V5 (#128)
Update V5
|
2024-11-08 19:46:50 +08:00 |
|
|
|
e4f1a4fe97
|
Update V4 version (#92)
* update train_lora && update deepspeed && update training with max token length
* fix bug in train.py
* fix bug in training_with_video_token_length
* update v2v && update v2v api
* add rope2d embedding precomputation; move text encoder to dataloader to reduce gpu memory consumpution
* add cuda multi-stream to speedup vae encode
* update new vae && new comfyui
* fix some bug in training code
* Add lcm lora (#89)
Co-authored-by: xuanyuan.lb <xuanyuan.lb@alibaba-inc.com>
* Update Training Code and fix bug in low vram mode
* fix bug in low vram mode
* update report
* update cfg
* actual text clip
---------
Co-authored-by: mengli.cml <mengli.cml@alibaba-inc.com>
Co-authored-by: liubo0902 <38622806+liubo0902@users.noreply.github.com>
Co-authored-by: xuanyuan.lb <xuanyuan.lb@alibaba-inc.com>
|
2024-08-19 11:22:17 +08:00 |
|
|
Bubbliiiing
|
616a35425d
|
Fix the issue where Lora training cannot load state, fix the issue where training cannot eval (#66)
* rename the comfyui files and new readme
* fix bug in eval and load lora state dict
|
2024-07-18 15:47:51 +08:00 |
|
|
bubbliiiing
|
ebe1a64a1c
|
Create Code
|
2024-04-12 17:42:52 +08:00 |
|