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
- [Perf] [NPU] dcp: DSA extend gathers the prefix KV when that moves fewer bytes
- [Perf] [AMD] deterministic: Allow AITER decode attention on gfx950
- [Fix] dsa: Route mixed-chunk batches to the RAGGED top-k transform for bf16-only sparse prefill
- [DSA] Support flashmla_sparse_q8 prefill backend on CUDA
- [Feat] [AMD] torch-profiler: Enable torch.profiler GPU tracing on torch 2.12 / ROCm 10.1
- [Perf] [AMD] Qwen3.8: Enable fused GDN packing for ratio-3 heads
- [Fix] qwen3_coder: Preserve structured values in qwen3_coder mixed unions
- [Fix] mm: encode a multimodal item once per request in chunked prefill
- [Feat] [AMD] dsv4: Dispatch Mega-mHC prefill to aiter on gfx950
- [Feat] hicache: Export host Mamba pool occupancy next to host KV
- Docs
- Python not yet supported