[INTEL_HPU] enable tensor_wise_fp8 kernels - QkvRope #2182

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

Open

yanfeich wants to merge 1 commit into PaddlePaddle:develop from yanfeich:fp8qkvrope

Collaborator

yanfeich commented Nov 18, 2025

fused_qkv_rope

combine bf16/fp8 as unique kernel
split fp8 qkv_projection and output fp8 q/k/v

paddle-bot bot commented Nov 18, 2025

Thanks for your contribution!

paddle-bot bot added the contributor label

LeoZhao-Intel approved these changes

View reviewed changes


          enable tensor_wise_fp8 kernels

6b9fe1d

yanfeich force-pushed the fp8qkvrope branch from a3ad832 to 6b9fe1d Compare

November 19, 2025 06:11

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels