Skip to content

Pull requests: fla-org/flash-linear-attention

New pull request New
55 Open 802 Closed
55 Open 802 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

[Fix] Support odd-dimensional chunk linear attention on Hopper
#1262 opened Sep 18, 2026 by taking-lying-flat Contributor Loading…
4 of 5 tasks
[CI] Include nested imports in find_dependent_tests import scan
#1260 opened Sep 18, 2026 by aniruddhaadak80 Loading…
4 of 5 tasks
[Attn] Use the varlen kernel for equal-length dense prefill bug Something isn't working needs-verification Lacks real execution evidence (CI skipped / no before-after data)
#1259 opened Sep 18, 2026 by arcusbuilds Loading…
4 of 5 tasks
1
[Fix] Return correct gradient count in PackSequenceFunction.backward bug Something isn't working
#1258 opened Sep 17, 2026 by aniruddhaadak80 Loading…
4 of 5 tasks
[Conv] Validate effective backend after FLA_CONV_BACKEND override in ShortConvolution bug Something isn't working
#1257 opened Sep 17, 2026 by aniruddhaadak80 Loading…
4 of 5 tasks
[Ops] Implement Triton backend for MomentumDeltaNet enhancement New feature or request
#1256 opened Sep 16, 2026 by bhaochen Contributor Loading…
4 of 5 tasks
[Conv] Add CUDA backend for contiguous context parallelism enhancement New feature or request
#1254 opened Sep 16, 2026 by Tytskiy Loading…
4 of 5 tasks
[Perf] Respect gradient demand in fused linear cross entropy performance
#1253 opened Sep 15, 2026 by ZenAlexa Loading…
4 of 5 tasks
1
[Perf] Optimize KDA fwd inter_solve_fused kernel for Ascend NPU ascend-npu Ascend NPU (triton_ascend) related performance
#1252 opened Sep 15, 2026 by qianjinqiu Contributor Loading…
4 of 5 tasks
1
[Fix] Fix NPU accuracy for unaligned dims via last-dim DMA padding and leftover masking ascend-npu Ascend NPU (triton_ascend) related bug Something isn't working
#1250 opened Sep 15, 2026 by sunyi0505 Contributor Loading…
4 of 5 tasks
1
15
[Fix] Skip unneeded student gradients in fused KL loss bug Something isn't working
#1248 opened Sep 13, 2026 by taking-lying-flat Contributor Loading…
4 of 5 tasks
1
[Perf] Reuse L2Norm tuning across packed lengths performance
#1247 opened Sep 13, 2026 by taking-lying-flat Contributor Loading…
4 of 5 tasks
1
[Test] Add regression coverage for token_shift cache on first single-token calls enhancement New feature or request
#1246 opened Sep 13, 2026 by fleeb83 Loading…
4 of 5 tasks
1
[NPU] Add single-device NPUGraph support for KDA ascend-npu Ascend NPU (triton_ascend) related
#1245 opened Sep 12, 2026 by FightingZhen Loading…
4 of 5 tasks
10
[Fix] Make GDN-2 safe_gate diagonal path independent of future gates bug Something isn't working
#1244 opened Sep 12, 2026 by xy200303 Contributor Loading…
4 of 5 tasks
[Perf] Reuse chunk metadata for graph-capturable state allocation performance
#1243 opened Sep 11, 2026 by ZenAlexa Loading…
4 of 5 tasks
3
[Fix] Make KDA safe_gate diagonal path independent of future gates needs-verification Lacks real execution evidence (CI skipped / no before-after data)
#1242 opened Sep 11, 2026 by xy200303 Contributor Loading…
4 of 5 tasks
1
[Fix] Fold chunk state launch grids to fit the 65535-block limit bug Something isn't working
#1239 opened Sep 10, 2026 by xy200303 Contributor Loading…
4 of 5 tasks
4
[Ops] Extend fixed-capacity CUDA Graph support for GDN and KDA enhancement New feature or request needs-verification Lacks real execution evidence (CI skipped / no before-after data) waiting-author Reviewer acted; ball is in the author's court
#1230 opened Sep 7, 2026 by 1418036869abc-cmd Loading…
4 of 5 tasks
2
[Refactor] Bind backend dispatch to local registries enhancement New feature or request
#1223 opened Sep 6, 2026 by yzhangcs Member • Draft
4 of 5 tasks
10
[Fix] Correct packed loss boundaries and cache defaults bug Something isn't working
#1217 opened Sep 3, 2026 by taking-lying-flat Contributor Loading…
4 of 5 tasks
4
Gate solve_tril dot precision on whether TF32 is faster, not whether it exists bug Something isn't working waiting-author Reviewer acted; ball is in the author's court
#1205 opened Aug 30, 2026 by arbi-dev Contributor Loading…
[Rotary] Separate varlen lengths from cache capacity bug Something isn't working
#1204 opened Aug 30, 2026 by taking-lying-flat Contributor Loading…
4 of 5 tasks
2
[GDN] Add configurable output gate activation enhancement New feature or request minor Low-value-density change (typo/docs/small validation), batch-process
#1182 opened Aug 27, 2026 by alifurkanstahl Loading…
4 of 5 tasks
[Docs] Add software health index badge minor Low-value-density change (typo/docs/small validation), batch-process
#1153 opened Aug 19, 2026 by Nayjest Loading…
5 tasks done
ProTip! Follow long discussions with comments:>50.