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] [AMD] dsa: Skip the indexer on eager idle ranks
- [Fix] disaggregation: Relayout DCP-sharded DSA draft KV with the target
- [Feat] [AMD] disaggregation: Allow DSA DCP decode behind DCP1 prefill
- [Perf] [AMD] dsa: Allow retaining GLM kv_b BF16 under DCP
- [Bug] ROCm absorb-BMM hand config aborts Triton compilation at M=256 on gfx950
- [Fix] dsa: Remove host syncs from rank-divergent eager paths
- [Bug] ROCm DCP-PD allocator reserve growth precedes HIP resource aborts under burst load
- [Perf] [AMD] qwen3: Use the fused QK-norm + RoPE + KV write in decode with an fp8 KV cache
- [Fix] [AMD] aiter: Use the layer's softmax scale in MHA prefill and decode
- [Perf] [NVIDIA] DFlash: CUDA Graph support on DFlash in DP Attention mode
- Docs
- Python not yet supported