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] gigachat35: Keep streamed tool_index contiguous and honor SGLANG_FORWARD_UNKNOWN_TOOLS
- [Fix] [AMD] dsv4: Skip retired splits in the HIP sparse decode split reduce
- [Fix] bench: account for native streaming tokens independently of visible text
- [Fix] k2_v3: Treat both Optional spellings alike when coercing arguments
- [Fix] inkling: Skip KV stores for CUDA-graph padding tokens in the fused attn prologue
- [Perf] [AMD] glm5_next: Allow Quark MXFP4 shared expert fusion at TP2
- [Bug] [NPU] PP + NEXTN produces garbled output with NPU Graph enabled on GLM-5.2
- [Fix] [NPU] graph: Fix PP tensor handling in speculative replay
- fix(npu): register every physical HiCache host buffer for the L3 store
- [Fix] quantization: Honor ignored_layers for INT8 MoE
- Docs
- Python not yet supported