Skip to content

Sparse decode attention: per-token reference lifecycle for async self-speculation - #157

Open
kunkunblueberry wants to merge 32 commits into
InferMatrixLab:mainfrom
kunkunblueberry:kunkun/rfc156-sparse-decode-async
Open

kunkunblueberry wants to merge 32 commits into
InferMatrixLab:mainfrom
kunkunblueberry:kunkun/rfc156-sparse-decode-async

sparse decode attention: per-slot contiguity, owned kernel scratch

f140ad5
Select commit
Loading
Failed to load commit list.
Sign in for the full log view

Annotations

1 warning and 1 notice
lint and CPU unit tests
succeeded Nov 5, 2026 in 2m 35s