Add Unsupported Languages to Base Model #3639
pourmand1376
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
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
Yesterday, I was talking to @andreaskoepf on discord about how to add a new language to Base LLM.
Today I saw this comment from @somerandomguyontheweb:
It seems that there are others like me who would like to fine-tune LLMs for unsupported languages like Persian.
This can be the place to discuss it. About asked question, I only know that he used this repository as the base and changes lots of things to make it work. I will ask him to give further details.
However, I think this repo can potentially serve as a repo for training base LLMs also.
I think we need a clear guide for people like me on how to do this thing. What I've seen so far, is that the Open-assistant team has done a great job for SFT fine-tuning. But there seems to be no code for fine-tuning base LLMs for other languages.
Beta Was this translation helpful? Give feedback.
All reactions