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] deepseek-v4.1: score the torch indexer chunk in place
- [wip][Feature] add eagle4 spec decode
- [Perf] QSA: Remove the D2D copy in prefill generated kv gather
- [Fix] deepseek-v41: Split thinking when SGLANG_DEFAULT_THINKING is set
- [Feat] glm5: Expose per-layer MLA key buffer and use_fused_topk in binding
- [Test] scheduler: Reset the published context after each prefill-adder test
- [Feat] disaggregation: Support shared Mooncake staging buffers for PD role switching
- [Bug] EPD encoder allocates unused pinned embedding pools on nonzero TP ranks
- [Bug] pause_generation(mode="abort") with --tokenizer-worker-num > 1 can let a request reach the scheduler after the KV cache is released
- [Fix] function_call: Stream Kimi K3 tool-call arguments incrementally
- Docs
- Python not yet supported