Skip to content

support fa3 quant for v0.9.1-dev #1695

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

Open
wants to merge 1 commit into
base: v0.9.1-dev
Choose a base branch
from

Conversation

22dimensions
Copy link
Contributor

@22dimensions 22dimensions commented Jul 9, 2025

  1. move fa3 quant from mindie turbo to vllm-ascend
  2. remove mindie turbo quantizer since all quantization method is inside vllm-ascend

Copy link

This pull request has conflicts, please resolve those before we can evaluate the pull request.

@22dimensions 22dimensions force-pushed the v0.9.1-dev-faquant branch 2 times, most recently from c7fd67c to 1d2e23e Compare July 12, 2025 08:33
quant_per_tensor)


class AscendFAQuantAttentionMethod:
Copy link
Collaborator

Choose a reason for hiding this comment

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

Does this feature currently affect accuracy?

@22dimensions 22dimensions force-pushed the v0.9.1-dev-faquant branch 10 times, most recently from 4b2faea to d7a3c30 Compare July 18, 2025 03:33
Signed-off-by: 22dimensions <waitingwind@foxmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants