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] models: use fused apply_qk_norm in Exaone, HunYuan-V3, AfMoE, and BailingMoE
- [Feat] scheduler: Correlate prefill admission scans with DP forward rounds
- [Fix] [NPU] loader: Preserve packed NZ weights in presharded checkpoints
- [Fix] mooncake: Surface transfer thread device binding failures to sender polls
- [Perf] [NPU] allocator: Share the Triton bound for one-block KV extends
- [Feat] model: Add Ripple-65B Preview (ricursiveAI/Ripple-65B-Preview) support
- [Perf] [AMD] deepseek-v4.1: Keep the exact decoder SWA tail in HIP breakable prefill graphs
- [Fix] modelslim: don't mutate the caller's quantization config
- [Fix] GLM: fold reasoning_effort onto the template's levels for tool grammars
- [Fix] [XPU] diffusion/z-image: Fix missing QK norm + RoPE fusion path for B=1 (already used for B>1)
- Docs
- Python not yet supported