Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

DPO/PPO/etc should default to HF chat template #458

Open
hamishivi opened this issue Nov 25, 2024 · 0 comments
Open

DPO/PPO/etc should default to HF chat template #458

hamishivi opened this issue Nov 25, 2024 · 0 comments

Comments

@hamishivi
Copy link
Collaborator

Currently, our codebase is somewhat hard-coded when it comes to chat templates. We should mainly assume the SFT stage correctly sets the chat template in the tokenizer (which it currently does) and just load from that in all post-SFT training stages. We can probably have a few alternate options if it makes sense.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@hamishivi