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
- [DeepSeek-V4] Forward auxiliary hidden states across PP stages
- fix: support bidirectional sliding window attention in torch native backend
- [Bug] Non-causal sliding-window attention is ignored in TorchNativeAttnBackend
- learn-basics
- [DSV4.1] Add Hopper block-FP8 path and tuned H20-3e configs
- [Bug] DeepSeek-V4 checkpoint with fp8 routed experts scores 10pp lower on AIME25 than vLLM (same weights, same node, same harness)
- [NPU] [DOC] Add Ascend NPU (Atlas 800I A3) deployment scenarios to the GLM-5.2 cookbook
- [Router] Report the tier held and block-0 presence from a prefix lookup (1/3)
- [Router] Expose cache-aware locality on /metrics: query / matched / selected blocks (2/3)
- [Router] Add the cache-locality rows to the Grafana dashboard (3/3)
- Docs
- Python not yet supported