LoRA Finetune of DocOwl

[xdaiycl]

I performed LoRA fine-tuning on mPLUG-DocOwl1.5-Stage1, but I couldn't find the corresponding inference script. When I use docowl_infer.py, it throws an error. Could you please provide a usable inference script for LoRA?

[Chunchunwumu]

Hello, have you solved this issue? I also meet it. Could you provide me the solution?

[dot-birdie]

Sounds potentially like the same problem in the load_pretrained_model function that I encountered previously:

refer to issue #123