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
- [DSA][DCP] HiCache host tier under decode context parallel (stacked on #39638)
- [Bug] SGLANG_FORWARD_UNKNOWN_TOOLS is not honoured by four tool-call detectors on the one-shot path
- [router] Record sgl_router_decode_affinity_total on the PD decode selection path
- [Perf][Logprob] Remove per-request CUDA syncs from token-id logprob path
- [CI][NPU] Stop running the NPU PR suite on every PR
- [AMD] Add JIT cooperative DSA top-k
- [MPS][MLX] Fix default memory budgeting and KV write retention
- [Bug] GitHub: Standard users do not have permission to reopen issues.
- [diffusion] model: support HiDream-O1-Image
- Regression on main — DFLASH speculative decoding with Qwen4-Exp crashes in decode CUDA-graph capture; the fix is the Qwen4-Exp capture guard in #38642
- Docs
- Python not yet supported