Skip to content

Pull requests: sgl-project/sglang

New pull request New
1,702 Open 12,846 Closed
1,702 Open 12,846 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 tool_choice stall on multi-call prompts (#17998)
#19848 opened Mar 4, 2026 by AjAnubolu Loading…
1
1
[misc] add env for http keep alive timeout
#19847 opened Mar 4, 2026 by happierpig Loading…
5 tasks
3
[quant] fix fp32 downcasting
#19844 opened Mar 4, 2026 by zhooooong Loading…
5 tasks
1
[AMD] Use bfloat16 for correction_bias in AITER FP8 path to avoid runtime dtype conversion for dsv3 deepseek run-ci
#19843 opened Mar 4, 2026 by inkcherry Loading…
5 tasks
2
[NPU][Feature] eliminate dsv3 redundant rotary embed calculation
#19842 opened Mar 4, 2026 by liupeng374 Loading…
5 tasks
1
[AMD][Bugfix] Fix get_global_server_args NameError on ROCm in fused_moe
#19840 opened Mar 4, 2026 by michaelzhang-ai • Draft
5 tasks done
1
[diffusion] runtime: introduce generation service facade diffusion SGLang Diffusion
#19839 opened Mar 4, 2026 by richl9 Loading…
5 tasks
1
[diffusion] docs: add diffusion runtime architecture boundary guide documentation Improvements or additions to documentation
#19838 opened Mar 4, 2026 by richl9 Loading…
5 tasks
2
[OPT]Skip the first delayer to maximize the BS of the decoding. run-ci
#19836 opened Mar 4, 2026 by chenxu214 Loading…
5 tasks
2
fix cuda graph capturing error in sm120 mxfp8 triton path
#19835 opened Mar 4, 2026 by wolfcomos Loading…
5 tasks
1
[AMD] CI - Add MI35x nightly tests for kv-cache-fp8 and allreduce-fusion (DeepSeek) amd deepseek
#19834 opened Mar 4, 2026 by yctseng0211 • Draft
5 tasks
1
Fix Qwen3.5 pipeline parallelism crash
#19833 opened Mar 4, 2026 by AjAnubolu Loading…
3 tasks
1
1
feature: support X-Data-Parallel-Rank header to specific dp-rank.
#19832 opened Mar 4, 2026 by doujiang24 Loading…
5 tasks
2
Add support for InstantTensor documentation Improvements or additions to documentation
#19830 opened Mar 4, 2026 by arlo-aisys Loading…
4 of 5 tasks
1
[AMD] aiter a8w8 gemm configuration
#19826 opened Mar 4, 2026 by seungrokj Loading…
5 tasks
2
[NPU][Bug fix] context parallel bug fix deepseek npu
#19820 opened Mar 4, 2026 by liupeng374 Loading…
5 tasks
1
Fix torch.compile/dynamo crash with Qwen3 QK-norm in piecewise CUDA g…
#19818 opened Mar 4, 2026 by jasperjiaguo Loading…
5 tasks
1
[diffusion][WIP] support realtime krea diffusion diffusion SGLang Diffusion
#19817 opened Mar 4, 2026 by IPostYellow • Draft
5 tasks
1
[FIX] Disable fused shared experts on pre-Hopper GPUs deepseek
#19816 opened Mar 4, 2026 by xueliangyang-oeuler Loading…
5 tasks
1
[AMD] Add bf16 MoE weights padding quant LLM Quantization
#19814 opened Mar 4, 2026 by Emmanuel0612 Loading…
5 tasks
3
[BUGFIX]Releasing redundant memory of w13_weight and nz when the FUSEEP feature is enabled run-ci
#19813 opened Mar 4, 2026 by chenxu214 Loading…
5 tasks
2
Fix Qwen3.5/Qwen3Next MTP EPLB compatibility
#19812 opened Mar 4, 2026 by AjAnubolu Loading…
5 tasks
1
[PD] Add graceful RDMA shutdown for Mooncake/NIXL disaggregation connectors
#19810 opened Mar 4, 2026 by Kangyan-Zhou • Draft
4 tasks
1
Add priority scheduling for PD disagg with a header support for x-sglang-request-priority. model-gateway run-ci
#19808 opened Mar 4, 2026 by harrisonlimh Loading…
1 of 5 tasks
2
Fix issue 19717 by making q0_indptr uniform strided instead of packed run-ci
#19807 opened Mar 4, 2026 by kpham-sgl Loading…
5 tasks done
2
ProTip! Mix and match filters to narrow down what you’re looking for.