Dao-AILab/flash-attention fa4-v4.0.0.beta19
on GitHub

pre-release6 hours ago

What's Changed

fix(hd256/sm100): make q/k/v contiguous before dedicated hd256 kernel by @yunweili3 in #2666
[Cute,Bwd,Sm100] add sparse MLA (Deepseek v4) backward kernels by @jayhshah in #2621

New Contributors

@yunweili3 made their first contribution in #2666

Full Changelog: fa4-v4.0.0.beta18...fa4-v4.0.0.beta19

Check out latest releases or
releases around Dao-AILab/flash-attention fa4-v4.0.0.beta19

Don't miss a new flash-attention release

NewReleases is sending notifications on new releases.

Get notifications