Is it possible to LoRA Fine Tune Mistral 7B after loaded using 8 / 4 bits with Transformer library?
#5755
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Tried FP16 but encounter CUDA out of memories. I couldn't find anyone discussing on this issue.
Beta Was this translation helpful? Give feedback.
All reactions