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
- [Feature] Model registry: eager import of all model modules hides failures (one bad dependency → 0 models registered) and adds ~2 s startup
- [Fix] dsa: Rebuild draft page table after DP-attention padding
- [Perf] [AMD] spec: Skip the vocabulary-wide sort in MTP rejection-sampling verify
- [Perf] [AMD] gdn: Add an aiter backend for linear-chain target verify
- [Perf] [AMD] deepseek-v4.1: Run the HIP sparse prefill on the FP8 KV cache
- [Fix] [AMD] dsv4: Let breakable prefill graphs capture without the SWA tail
- [Feat] [AMD] diffusion: Add MXFP6 and A6W4 online GEMM quantization
- [Fix] kimi-k3: Route DFLASH target verification to cutedsl_mla under DCP
- [Fix] openai: Avoid duplicate BOS on text-only requests to multimodal models
- [Fix] rust-server: Return named and required tool calls in non-streaming responses
- Docs
- Python not yet supported