can we use this model to finetune it on our specific dataset , like how other models hosted on hugging face is done.
#26
by
MukeshSharma
- opened
Describe ho that can be done
Has there been any progress on this?
Yes, it is a 30b LLaMA model, it can be fine-tuned. the GH repo has some more in-depth instructions on this if you want to do OA RLHF, but otherwise you can finetune it like anything else.