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] trtllm_mha: Pack a graph-built page table for eager reads
- [Feat] observability: Add request-level frontend wall-time histograms
- [Perf] sglang-kernel FA3 picks MmaPV_is_RS=false for hdim 128 non-causal on H200 (1.5-3% slower than flash-attn's hopper build)
- [Feat] speculative: Support synchronous STANDALONE PD handoff
- [Fix] [AMD] diffusion: install amd-smi so FLUX.3 serves on ROCm on gfx1151
- [Fix] Log: Skip leaked user input fields (e.g. video_config._question) via nested field skip support
- [RFC] Agent harness x SGLang: keep multi-turn agent prefixes cacheable
- [Perf] hicache: Extend prefix_keys in place across storage batches
- [Feat] multimodal: Add bounded GPU feature publication admission
- [Fix] spec: Honor deterministic inference in draft and verify softmax
- Docs
- Python not yet supported