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] compressed-tensors: Reject WNA16 MoE TP shards that split a scale group
- [Fix] [AMD] custom-all-reduce: Support expandable segments
- [Perf] kimi-k3.1: Share ViT work across prefill CP ranks
- [Feat] disagg: Admit SWA-compress models such as Inkling in the PD decode radix cache
- [Perf] rotary_embedding: Avoid tail slicing and redundant scans in bailing_3drope_get_input_positions_tensor
- [Perf] [AMD] mla: Read the fused_qkv_a slices in place in the gfx1250 q_a / kv_a RMSNorm, one launch
- [Perf] deepseek-v4.1: Fuse vision Q/K rotary on CUDA
- [Fix] disagg: Notify decode when prefill aborts a request
- [Fix] eplb: Reject expert distribution record calls when the recorder is disabled
- [Fix] [NVIDIA] moe: Fall back for strided router logits
- Docs
- Python not yet supported