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] openai: keep the tool-call parser on under tool_choice="none"
- [Feat] [AMD] mamba: Store hicache checkpoints without a GPU slot
- [Feat] [NPU] speculative: Allow block verification for EAGLE/EAGLE3/NEXTN chain path
- [RFC] scheduler: Overlap prefill scheduling and KV loading with model execution
- [Fix] detokenizer: Use chunk-relative read offsets after state eviction
- [Feat] scheduler: Overlap prefill scheduling and KV loading with model execution
- [Feat] [AMD] vision: Support ViT CUDA graph with the aiter_attn backend
- [Docs] [AMD] Qwen3.8-27B cookbook: Add an MI355X MXFP4 recipe
- [Feat] [AMD] DeepSeek-V4.1: Prefill CP fixes on top of #43465: zero-row groups, decode attention TP, image inputs
- [Fix] DeepSeek-V4.1: Keep DP-attention decode collectives off per-graph registration on ROCm
- Docs
- Python not yet supported