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
- [Perf] [NVIDIA] speculative-decoding: Bind generic provider strategies
- [Fix] [SM120] dsv4: Keep the V4.1 indexer usable on SM12x
- [Fix] [NPU] dspark: Avoid double-counting verify KV lengths in graph replay
- [Perf] dsa: Elide shared-layer index-K under HiSparse decode to raise batch size
- [Test] [NPU] glm5_2: Add 2P2D PD-sep gpqa cases and unify evalscope retry loop
- [Fix] multimodal: Share typed identities for Kimi prepared inputs
- [Bug] On Python 3.13, installing sglang 0.5.22 requires a Rust toolchain (hidden, undocumented, caused by outlines==0.1.11 → outlines_core==0.1.26 exact pin)
- [Perf] diffusion: replay LTX head-gate and RMSNorm launches
- [Doc] [NVIDIA] cookbook: Recommend the CuTe DSL all-reduce fusion backend for GLM-5.2 GB300 TP4 low latency
- [Fix] openai: Report server context length in /tokenize max_model_len
- Docs
- Python not yet supported