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] Handle JSON Schema union param types in MiMo tool call detector
- [ROCm] gfx950: schedule the fused sparse-MLA kernel, gate the MLA q-absorb, and fuse the EAGLE draft proposal
- [Bugfix] Falcon-H1: do not .float() tied LM head embeddings in place
- [Bug][Simulator] SGLang Simulator cannot load hybrid GDN (qwen3_5) models: the CPU engine requires sgl_kernel CPU ops the CUDA wheel does not ship
- [Diffusion] Force VAE tiling on for Qwen-Image 2.1 on gfx1151
- Fix XQA draft extend with FP8 KV cache in trtllm_mha
- Patch 21
- [AMD][Perf][MiniMax-M3] Route sparse attention path to use Aiter's `fused_qknorm_idxrqknorm`
- [AMD][Perf][MiniMax-M3] Fuse the bf16->fp8 main-KV cast into the sparse-layer paged scatter
- [AMD][Perf][mem_cache] Cut three per-step kernel launches from paged decode alloc
- Docs
- Python not yet supported