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] kimi-k3: Optimize single-token MoE routing and finalize
- [Perf] HiCache PP-sync drain makes the PP prefill scheduler lock-step across ranks
- [Perf] [AMD] Qwen4-Exp (Qwen3.8-Flash-Next): FlyDSL monokernel for speculative verify / MTP draft on MI355X
- [Fix] offloader: Restore forward wrappers after exceptions
- [Bug] PD prefill PP8 + HiCache cache mode hangs: PP0–PP3 blocked in the HiCache PP-sync drain
- [Fix] srt: Replace bare except: with except Exception:
- [Doc] cookbook: add Muse Spark unified memory control
- [Perf] beam: Add optional cuTile BF16x3 router GEMM
- [Doc] lora: document PEFT adapter compatibility and fused target modules
- [Fix] anthropic: Accept thinking.type=enabled without budget_tokens
- Docs
- Python not yet supported