sglang
https://github.com/sgl-project/sglang
Python
SGLang is a high-performance serving framework for large language models and multimodal models.
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported1 Subscribers
Add a CodeTriage badge to sglang
Help out
- Issues
- [multimodal] Add per-processor default image count limit
- [NPU] Scope nightly to Kimi-K3 W4A8 32p accuracy test (gpqa) with cann9.1.0-a3-20260908 image
- fix(rocm): support unified KV in DSV4 multi-stream
- [AMD] Prevent DeepSeek-V4 MTP GPU faults during graph replay
- [Bug] 主线版本ascend后端对swa模型的注意力计算有bug
- ci(xpu): card-level LPT sharding launcher for stage-a job
- [Feature] Expose multimodal preprocessing latency metrics (media download / load / processor) in Prometheus
- docs(cookbook): add Ascend NPU (Atlas 800I A3) deployment scenarios for Qwen3.6
- [SpecDecode] Add opt-in compact TP4 linear target verification
- [Feature] Add flashinfer_b12x NVFP4 backend and tiny_gemm decode path for SM120/SM121 (RTX 5000)
- Docs
- Python not yet supported