Skip to content

[feat] support fa3 backend for pd disaggregated #2695

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

Merged
merged 5 commits into from
Jul 3, 2025

Conversation

yuanlehome
Copy link
Collaborator

@yuanlehome yuanlehome commented Jul 3, 2025

  1. Currently, it is only supported for use in PD, by export FD_ATTENTION_BACKEND=FLASH_ATTN
  2. Currently, We need skip profile, by specifying --num-gpu-blocks-override
  3. By the way, delete the use_fast_ffn parameter

Copy link

paddle-bot bot commented Jul 3, 2025

Thanks for your contribution!

@yuanlehome yuanlehome force-pushed the add_fa3_backend_v2 branch from 0f83e81 to f8157ef Compare July 3, 2025 12:14
Copy link
Collaborator

@EmmonsCurse EmmonsCurse left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@yuanlehome yuanlehome merged commit 240bdac into PaddlePaddle:develop Jul 3, 2025
2 of 3 checks passed
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

Successfully merging this pull request may close these issues.

2 participants