How to evaluate LLM fine tuning? #736
Stealthwriter
started this conversation in
General
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
After fine tuning, I want to evaluate the new fine tuned LLM, how to know if its accurate? training loss doesn't give you a clear answer.
I spend a lot of time manually evaluating the model, is there an alternative to that?
Beta Was this translation helpful? Give feedback.
All reactions