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
- [HiCache] Continue eviction after a failed file unlink
- [Fix] hicache: Carry Mamba slot side states through zero-copy storage backends (Qwen3.8-Flash-Next on Mooncake/NIXL)
- [Fix] hicache: Fix NIXL hybrid bounce buffers: copy page bytes exactly and pad slots for O_DIRECT
- [AMD] Fuse the DeepSeek-V4 shared expert by default on gfx950 with the aiter MoE and MXFP4 experts
- [AMD] Quantize the DSv4 indexer FP4 query with one program per token and 16-head group
- [AMD] Fuse MXFP8 quant into the small MoE sort
- [AMD] DeepSeek-V4.1: route decode MXFP8 dense GEMMs above 8 rows to aiter on gfx950
- [AMD] DeepSeek-V4.1: faster prefill MXFP8 dense GEMM at large M on gfx950
- [AMD] DeepSeek-V4.1: TP4 A8W4 MoE rows in AITER (pointer)
- [AMD] DeepSeek-V4.1: run the Engram lookup all-reduce on the custom all-reduce path on gfx950
- Docs
- Python not yet supported