forked from sgl-project/sglang
-
Notifications
You must be signed in to change notification settings - Fork 0
Pull requests: character-tech/sglang
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
triton attention: mixed-batch split dispatch (decode rows -> decode kernels) (reorg 6/9)
#6
opened Aug 26, 2026 by
mike-char
Loading…
triton attention: sync-free SWA metadata, per-shape launch-config table, split-KV policy (reorg 5/9)
#5
opened Aug 26, 2026 by
mike-char
Loading…
generic features: fast top-k/p sampling, --disable-multimodal, HiCacheFile page-cache control, allocator guard (reorg 4/9)
#4
opened Aug 26, 2026 by
mike-char
Loading…
hicache hybrid-SWA: re-entrancy, ownership rematch, watermark eviction, ack-reap fixes (reorg 3/9)
#3
opened Aug 26, 2026 by
mike-char
Loading…
quantization/model correctness: fp8 padding, CT MoE padding, PTPC toggle, gemma-4 gelu_tanh (reorg 2/9)
#2
opened Aug 26, 2026 by
mike-char
Loading…
upstream backports: hicache/SWA picks + #35017 + ports of open PRs #33988/#22615 (reorg 1/9)
#1
opened Aug 26, 2026 by
mike-char
Loading…
ProTip!
Exclude everything labeled
bug with -label:bug.