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] [AMD] mori: Read MORI EPv2 AsyncLL segmented receive in place, skipping the receive copy
- [Feat] models: add public registration for runtime classes
- [Feat] [NPU] quant: Serve serialized block-FP8 MoE weights as MXFP8 on Ascend A5
- [Perf] sparsity: Vectorize BaseSparseAlgorithmImpl.retrieve_topk across active requests
- [Doc] hicache: Add recipe for attributing cache misses and HiCache restores via --kv-events-config
- [Bug] Mixtral, Llama 4 and FP8 Qwen3-MoE: a router stored in FP8 is loaded without its scale, with no error
- [Bug] sm_120 (RTX 50-series consumer Blackwell) cannot run any model out-of-the-box with CUDA 13 wheel install
- [Fix] minicpm5: Update the MiniCPM5 XML tool detector
- [Feat] comm: Add multicast symmetric-window transfers
- [Feat] sampling: Honor per-request sampling_seed without deterministic inference
- Docs
- Python not yet supported