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] scheduler: Skip --prefill-decode-interval when nothing is decoding
- [Perf] model_loader: Fault mmap'd safetensors views in on the CPU on integrated GPUs
- [Feat] [AMD] minimax: gfx950 sparse attention, NVFP4 KV and DSpark verify kernels
- [Feat] linear-attn: Bole spec-verify for GDN (drop per-draft SSM snapshots)
- [Fix] embedding: Force full recompute when pooling reads inside a cached prefix
- [Fix] [MUSA] fa3: Preserve draft-extend prefix KV attention
- [Feat] [AMD] mori: add XGMI data plane for single-node PD
- [Fix] [XPU] spec: Start NEXTN/MTP drafts on XPU and stop silent greedy verify
- [Perf] disagg/nixl: Opt-in warm-up transfer to each decode peer at registration
- [Bug] Mooncake PD disaggregation: DSA_TAIL raises "DSA tail slot index missing" when Prefill/Decode seq_len diverge on last_chunk
- Docs
- Python not yet supported