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

[Feature Request] expose unidirectional (causal) attribute of GQA #23409

Open
hann-wang opened this issue Jan 17, 2025 · 0 comments
Open

[Feature Request] expose unidirectional (causal) attribute of GQA #23409

hann-wang opened this issue Jan 17, 2025 · 0 comments
Labels
feature request request for unsupported feature or enhancement

Comments

@hann-wang
Copy link
Contributor

hann-wang commented Jan 17, 2025 •
edited
Loading

Describe the feature request

We have a unidirectional attribute in MHA, but it is missing in GQA because most LLMs are causal. Transformer-based Text-to-Image models, on the other hand, are not causal. We should expose this attribute in GQA to help facilitate the deployment of Text-to-Image models.

Describe scenario use case

Text-to-Image generation models (like Stable Diffusion 3) require attention implemented with causal=False.

PR

#23412
@hann-wang hann-wang added the feature request request for unsupported feature or enhancement label Jan 17, 2025
@hann-wang hann-wang changed the title [Feature Request] expose is_unidirectional (causal) attribute of GQA [Feature Request] expose unidirectional (causal) attribute of GQA Jan 17, 2025
@hann-wang hann-wang mentioned this issue Jan 17, 2025
Draft
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
feature request request for unsupported feature or enhancement
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@hann-wang