(Translated by https://www.hiragana.jp/)
vllm implementation · Issue #31 · vgel/repeng · GitHub
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.

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

vllm implementation #31

Closed
raywanb opened this issue Apr 17, 2024 · 4 comments
Closed

vllm implementation #31

raywanb opened this issue Apr 17, 2024 · 4 comments

Comments

@raywanb
Copy link

raywanb commented Apr 17, 2024

I'm trying to implement control vector into vllm codebase for mixtral model, but I was wondering where should I add the control vector to the layer. Should it be added before attention, fully connected, or after? Thanks @vgel

https://github.com/vllm-project/vllm/blob/a53222544c6385ee314a26fdf42eb14f5b4e5ad9/vllm/model_executor/models/mixtral.py#L270

@vgel
Copy link
Owner

vgel commented Apr 18, 2024

Awesome, very excited for VLLM support! The control vector gets added after the MLP/FC layer. However, I noticed you linked to Mixtral, not sure if that was intentional but FYI (I should note this more clearly in the README) control vectors don't work on MoEs currently, probably best to implement for dense models only.

@raywanb
Copy link
Author

raywanb commented Apr 27, 2024

Ok, got it thanks. btw is there a proper way of saving control vectors so ppl can upload them onto HuggingFace?

@rexarski
Copy link

rexarski commented May 8, 2024

Ok, got it thanks. btw is there a proper way of saving control vectors so ppl can upload them onto HuggingFace?

This might help! #23

@vgel
Copy link
Owner

vgel commented May 24, 2024

@raywanb

Ok, got it thanks. btw is there a proper way of saving control vectors so ppl can upload them onto HuggingFace?

There's an official way now, see #34 !

@raywanb raywanb closed this as completed Aug 8, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants