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
- [Feat] muse-spark: Infer hardware launch defaults and simplify cookbook recipes
- [Fix] scheduler: Fire priority preemption when KV runs out, and keep retracted KV behind a flag
- [Fix] benchmark: Keep the Mooncake trace's 512-token blocks in the mooncake dataset
- [Feat] scheduler: Add a skip-join MLFQ policy with cache-aware placement and a prefill gate
- [Feat] rust: deepseek-recipe chat encoding for sglang-processor, sgl-router and sglang-server
- [HiSparse V2][1/n] Support shared KV sources in HiSparse MLA swap kernels
- [Fix] [AMD] smallm_fp8_blockscale_gfx950: import renamed hipcc helpers
- [Feat] parser: Source-certified marker matching contract (P1, stacked on #45587)
- [Bug] Kimi-K3 on SM90: `flashinfer_mxfp4` silently computes GPT-OSS SwiGLU instead of SiTU
- [Feat] muse_spark: Pipeline parallelism with DFLASH speculative decoding
- Docs
- Python not yet supported