Skip to content

Pull requests: MoonshotAI/FlashKDA

New pull request New
10 Open 7 Closed
10 Open 7 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Handle all-empty varlen batches
#32 opened Aug 17, 2026 by 0z5a Loading…
Add reusable workspace support to the Python API
#31 opened Aug 17, 2026 by 0z5a Loading…
1
Add intermediate states and exact CP transition prototype
#30 opened Aug 14, 2026 by 0z5a Loading…
Reuse tile prefix in recurrence kernel
#29 opened Aug 14, 2026 by 0z5a Loading…
Add CUDA training kernels (fwd+bwd) for KDA
#28 opened Aug 8, 2026 by xy200303 Loading…
5
Add differentiable CPU backend for KDA research
#27 opened Jul 29, 2026 by xy200303 Loading…
1
Report CUTLASS setup and CUDA launch-configuration failures
#19 opened Jul 27, 2026 by morluto Loading…
1
Validate the independent FLA 0.5 compatibility path
#18 opened Jul 27, 2026 by morluto Loading…
Parallelize partial output-tile stores across the store warp
#17 opened Jul 27, 2026 by morluto Loading…
1
Tolerate one-ULP BF16 variation in correctness tests
#16 opened Jul 27, 2026 by morluto Loading…
ProTip! Adding no:label will show everything without a label.