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] mem_cache: Hybrid SWA HiCache fixes and --hicache-swa-host-ratio
- [Perf] [AMD] dsa: Default the DSA KV cache to fp8_e4m3 on gfx950
- [Fix] parser: Fail a DeepSeek-V4/V4.1 thinking turn that stops before </think>
- [Doc] golang: point the Basic Usage TOC entry at the heading that exists
- [Fix] cuda_graph: Disable full prefill CUDA graph for hybrid linear attention models (#48593)
- [Fix] utils: Wake waiting readers when a waiting writer is cancelled in `RWLock`
- [Fix] gdn: Add opt-in FP32 prefix-cache snapshots
- [Fix] diffusion/realtime: Keep the session open when a client sends a text frame
- [Fix] diffusion/realtime: Close the connection with 1011 when generation fails
- [Fix] diffusion/realtime: Tell clients why a request or event was rejected
- Docs
- Python not yet supported