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
- [Feat] HiSparse: Support DeepSeek-V4 with MTP and DSpark
- [HiCache] Support partial transfer groups for GLM-5.3
- [Fix] DeepSeek-V4.1: keep the mHC next-layer link out of the module tree
- [RFC] MoE routing simulator for prefill imbalance profiling: balanced reference and GPU-faithful replay
- [Fix] HiCache: report storage clear failure accurately
- [Feat] model-loader: Add model-aware checkpoint shard selection for Qwen MTP
- [Feat] model-loader: Enable model-aware Qwen MTP loading in RunAI ModelStreamer
- [Fix] lora: support PEFT rsLoRA and GDN adapters on Qwen3.5-family hybrid MoE
- [Feat] [NPU] speculative: Support --speculative-draft-window-size/-sink-size via draft-decode KV mask
- [Feat] valen: add native multimodal shared-state decisions
- Docs
- Python not yet supported