-
Notifications
You must be signed in to change notification settings - Fork 163
Pull requests: sgl-project/sgl-kernel-npu
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
[fix] Fix alltoall strategy dispatch TypeError when quant_mode is None
#637
opened Jul 29, 2026 by
1329009851
Contributor
Loading…
4 tasks done
feat: add NPU primitives for sparsity-driven KV offloading
#636
opened Jul 29, 2026 by
nzbej
Loading…
5 tasks done
fix(build): skip attentions build for DeepEP targets
#627
opened Jul 27, 2026 by
WenhaoGe
Contributor
Loading…
4 tasks done
feat(hicache): add async Mamba state transfers
#626
opened Jul 27, 2026 by
qybnb
Loading…
4 tasks done
[fix] Restore #600 quantization calling convention for backwar
#624
opened Jul 27, 2026 by
1329009851
Contributor
Loading…
4 tasks done
[refactor] use init_moe_routing instead of npu_moe_token_permute
#611
opened Jul 22, 2026 by
zzx-study
Contributor
Loading…
4 tasks
fix(ci): isolate multi-node cache path and k8s job per run+version
#607
opened Jul 22, 2026 by
EdwardXuy
Contributor
Loading…
[WIP] [feat] A5 deep fused moe supported
#602
opened Jul 21, 2026 by
666syh
Contributor
Loading…
4 tasks done
causal_conv1d kernel improvement using PTO kernel & benchmark results
#601
opened Jul 20, 2026 by
ChristosMatzoros
Contributor
Loading…
[WIP] Process k/v dimension in blocks to fit UB
#597
opened Jul 13, 2026 by
iridiumine
Contributor
Loading…
[DeepEP] increase MAX_K_VALUE_A2 from 10 to 16 to support topk > 10
#593
opened Jul 9, 2026 by
Yuxin1999
Loading…
Add sigmoid activation support to the NPU layernorm gated kernel
#578
opened Jun 30, 2026 by
ylyhyqsl
Loading…
[Attentions][Fix] Rename block sparse attention
#571
opened Jun 25, 2026 by
Napkin-AI
Contributor
Loading…
Previous Next
ProTip!
Add no:assignee to see everything that’s not assigned.