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
- [Feat] spec: DFlash adaptive verification integration (do not merge)
- [Perf] mamba: write the deduplicated conv verify window as one tile
- [Feat] gdn: run DFLASH and DSPARK verify on the ReplaySSM spec ring
- [Perf] spec: radix top-k for speculative verify target probabilities
- [Feat] spec: online acceptance estimator for drafts without a confidence head
- [Perf] gdn: fold ReplaySSM compact history in one product per head group
- [Feat] gdn: run target verify on packed ragged tokens (ReplaySSM)
- [Feat] spec: confidence-scheduled compact verify for DFLASH
- [Feat] hisparse: host KV tier for the QSA (Qwen4) sparse-attention backend
- [Fix] modelopt_fp4: skip swizzled blockscales the marlin runner never reads
- Docs
- Python not yet supported