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
- [Fix] [XPU] spec: Give EAGLE drafts triton attention when intel_xpu lacks their head size
- [Fix] [XPU] spec: Sample EAGLE verification on XPU instead of forcing argmax
- [Perf] [XPU] distributed: Default oneCCL to its Arc all-reduce on Arc GPUs
- [Perf] [XPU] qwen3: Fuse q/k RMSNorm and RoPE on XPU
- [Fix] [XPU] benchmark: Run the fused MoE Triton tuner on XPU
- [Perf] DeepSeek chat templates: concatenate tool text once instead of per-tool
- [Perf] sglang-api-types: Decode flat id lists directly instead of buffering them as Values
- [Fix] docker: Pin sgl-nvidia-mega to 9a21cb0 for megakernel API version 2
- [Perf] Cache schema type normalization and defer the template tool dump
- [Fix] disaggregation: Drop a decode's cached bootstrap infos when its prefill re-registers
- Docs
- Python not yet supported