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
- [mem_cache][PD] Fix decode radix cache ownership and allow EAGLE/EAGLE3 with decode radix cache
- [Unified Cache] Abort requests whose external linker load fails instead of crashing the scheduler
- [Bug] trtllm_mha for both prefill and decode on H200 (SM90) is accepted in v0.5.20 and returns wrong completions (gpt-oss-120b); v0.5.17 refused it; either half alone is correct
- [diffusion] Support MiniMax-H3 disaggregated serving
- [Feature] EFA SGLang image should work out of the box on AWS NVIDIA GPUs (NIXL & Mooncake)
- [Bugfix] GigaChat3 streaming: stop leaking protocol markers, keep arguments valid
- Fuse temperature scaling and softmax in the flashinfer sampler path
- [Perf] Elide redundant DSA index-K storage with HiCache L2
- [Bug][MLX] return_logprob silently returns empty logprobs without --mlx-enable-sampling, so /v1/score answers 500
- [MLX] Reject output logprobs without sampling
- Docs
- Python not yet supported