llm-course icon indicating copy to clipboard operation
llm-course copied to clipboard

Published 20 hours ago • verified publisher icon mlabonne

DPO with Axolotl

Open JohanWork opened this issue 1 year ago • 3 comments

It is possible to perform DPO with Axolotl. If I were to create a notebook for DPO fine-tuning, do you think it would be suitable for your repository?

JohanWork avatar Feb 15 '24 20:02 JohanWork

Hi Johan, thanks for your suggestion. I actually made one, will release it soon. Feel free to suggest improvements if you're interested, it's not perfect but it works haha.

On Thu, Feb 15, 2024, 20:09 JohanWork @.***> wrote:

It is possible to perform DPO with Axolotl. If I were to create a notebook for DPO fine-tuning, do you think it would be suitable for your repository?

— Reply to this email directly, view it on GitHub https://github.com/mlabonne/llm-course/issues/48, or unsubscribe https://github.com/notifications/unsubscribe-auth/ATL5EGUCTEEUJUIC2BRK62LYTZTOPAVCNFSM6AAAAABDK4LN4WVHI2DSMVQWIX3LMV43ASLTON2WKOZSGEZTOMZVHA2TSNI . You are receiving this because you are subscribed to this thread.Message ID: @.***>

mlabonne avatar Feb 15 '24 20:02 mlabonne

aa nice, looking forward to it. Will do!

JohanWork avatar Feb 15 '24 20:02 JohanWork

Released it here: https://twitter.com/maximelabonne/status/1759222499131199788 :)

mlabonne avatar Feb 19 '24 14:02 mlabonne