Cong Wan
Cong Wan
Also, is there a way to perform full fine-tuning instead of lora?
> Also interested in distributed training config Have u found a way for it, bro
yes, no problem. Go inside ./ai-toolkit-main/toolkit/buckets.py file, edit the resolutions_1024 list and you can simply customize your resolution.
yes, no problem. Go inside ./ai-toolkit-main/toolkit/buckets.py file, edit the resolutions_1024 list and you can simply customize your resolution.
I encountered a similar issue. It appears that the error arises when loading the transformer files. ---------------------------- 9.6 update Ok, I have solved it by "pip install accelerate"