Sitelet https://github.com/sgl-project/sglang/pulls
Skip to content

Pull requests: sgl-project/sglang

New pull request New
4,090 Open 24,651 Closed
4,090 Open 24,651 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

[Diffusion] Bound reusable Ulysses A2A staging buffers across shapes diffusion SGLang Diffusion run-ci
#36327 opened Aug 25, 2026 by BBuf Collaborator Loading…
[NPU] [DOC] Update CANN version in NPU installation docs documentation Improvements or additions to documentation
#36325 opened Aug 25, 2026 by amote-i Contributor Loading…
5 tasks
[Spec] Enable draft-extend CUDA graph on FlashAttention (fa3) via backend capability blackwell SM100/SM120 deepseek mthreads
#36324 opened Aug 25, 2026 by zkyue Contributor Loading…
3 tasks done
[diffusion] advise the kernel ahead of the courier on mapped layers diffusion SGLang Diffusion run-ci
#36323 opened Aug 25, 2026 by mickqian Collaborator Loading…
[diffusion] perf: fuse tanh-GELU into the LongCat-Image DiT FFN up-proj diffusion SGLang Diffusion
#36322 opened Aug 25, 2026 by ITerydh Contributor Loading…
5 tasks done
1
[Fix] Honor metrics content negotiation in gRPC sidecar
#36321 opened Aug 25, 2026 by TwistsOfFate Loading…
5 tasks done
[Diffusion] Fix MiniMax H3 WebUI inference settings diffusion SGLang Diffusion documentation Improvements or additions to documentation
#36320 opened Aug 25, 2026 by wplf Contributor Loading…
[diffusion] feat: support out-of-tree platforms and hook plugins diffusion SGLang Diffusion documentation Improvements or additions to documentation
#36319 opened Aug 25, 2026 by liquanfeng Contributor Loading…
4 of 5 tasks
[Diffusion] Add MiniMax-H3 packed varlen path for sla_attn backend diffusion SGLang Diffusion jit-kernel
#36318 opened Aug 25, 2026 by niehen6174 Contributor Loading…
1 of 6 tasks
[HiCache] Keep auxiliary load-back out of Full KV pending ownership run-ci unified-radix-cache
#36317 opened Aug 25, 2026 by weireweire Contributor Loading…
4
[compile][Spec] Let the EAGLE target keep its prefill CUDA graph on tc_piecewise blackwell SM100/SM120
#36316 opened Aug 25, 2026 by Duyi-Wang Collaborator • Draft
Require exactly one generation input type
#36315 opened Aug 25, 2026 by taylorty Loading…
3 of 4 tasks
[Spec][DSA] Scope --dsa-topk-backend to the target model deepseek documentation Improvements or additions to documentation run-ci
#36313 opened Aug 25, 2026 by zianglih Contributor Loading…
4 of 5 tasks
1
[Fix] Apply Bailing gemm1_clamp_limit on the FlashInfer CUTLASS MXFP4 path amd blackwell SM100/SM120 deepseek documentation Improvements or additions to documentation jit-kernel npu quant LLM Quantization sgl-kernel
#36312 opened Aug 25, 2026 by JustinTong0323 Collaborator Loading…
[NPU] [diffusion] Add support for HunyuanImage-3.0-Instruct diffusion SGLang Diffusion Multi-modal multi-modal language model
#36311 opened Aug 25, 2026 by e-martirosian Contributor • Draft
5 tasks
[Scheduler] Add cache-hit-aware over-admission for bounded Mamba requests documentation Improvements or additions to documentation memory-pool
#36310 opened Aug 25, 2026 by thunguo Loading…
5 tasks done
1
2
[AMD][Bugfix] Skip invalid fused MoE reduction for direct top-1 output
#36309 opened Aug 25, 2026 by bingxche Collaborator • Draft
5 tasks done
[CI][AMD] Limit HiCache MGSM eval concurrency on ROCm hicache Hierarchical Caching for SGLang
#36308 opened Aug 25, 2026 by bingxche Collaborator • Draft
3 of 4 tasks
[AMD][CI] Stabilize PyTorch sampling backend test on ROCm
#36307 opened Aug 25, 2026 by bingxche Collaborator • Draft
3 tasks done
1
[AMD] Pass USE_PDL explicitly in the fused MoE gate jit-kernel
#36306 opened Aug 25, 2026 by bingxche Collaborator • Draft
4 of 5 tasks
1
[Perf][DSA] Unify q8kv8 sparse-prefill topk_length source with the bf16 path; delete backscan kernel jit-kernel
#36305 opened Aug 25, 2026 by zkyue Contributor Loading…
[Kimi-K3] Enable fused KDA decode for TP4 jit-kernel
#36304 opened Aug 25, 2026 by foraxe Contributor Loading…
[diffusion] Support batched Cosmos3 action generation. diffusion SGLang Diffusion
#36301 opened Aug 25, 2026 by FxxxxU Loading…
2 of 5 tasks
[Weight Cache] Make daemon socket/ready paths configurable via env run-ci run-ci-extra
#36299 opened Aug 25, 2026 by liusy58 Collaborator Loading…
5 tasks
@liusy58
Fix/whisper xpu varlen encoder decoder
#36298 opened Aug 25, 2026 by SKRohit Contributor Loading…
ProTip! Mix and match filters to narrow down what you’re looking for.