Skip to content

[Model] Accelerate the Qwen2.5-Omni audio encoder: native packed attention + opt-in torch.compile and CUDA graph capture - #1

Closed
pratapyash wants to merge 725 commits into
review-base-omni-audio-encoderfrom
feat-upstream-omni-audio-encoder
Closed

[Model] Accelerate the Qwen2.5-Omni audio encoder: native packed attention + opt-in torch.compile and CUDA graph capture#1
pratapyash wants to merge 725 commits into
review-base-omni-audio-encoderfrom
feat-upstream-omni-audio-encoder

test(qwen2_5_omni): assert the packed qkv k-bias contract at construc…

bdbc4f4
Select commit
Loading
Failed to load commit list.

Select a check to view from the sidebar