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] srt: Promptly abort running requests on client disconnect (#46517)
- [Roadmap] TDX/CC Prefill–Decode Disaggregation
- [Feat] model-runner: Add an opt-in step-shape log
- [Fix] eplb: Warm default-group P2P transports on multi-node layouts
- [Perf] mega_moe: Raise BLOCK_M for decode-graph FP8xFP4 calls
- [Perf] [NVIDIA] minimax-sparse: Run the decode indexer CP on Hopper at TP4 and TP8
- [Feature] Support reasoning and content continuation through model renderers/encoders
- [Fix] diffusion: LingBot QKV sharing, causal RoPE on no-cache path, uneven split guard
- [Perf] diffusion model: Accelerate conditioning hashing with GPU BLAKE3
- [Feat] [NPU] hicache: Support layer-first split host transfers
- Docs
- Python not yet supported