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] hicache: Enable validated dynamic storage backends with DCP
- [Fix] encoder: allocate embedding cache resources only on rank 0
- [Feat] attention: Block-bidirectional (prefix-LM) attention for decision models via bidir_prefix_len
- [Feat] [CPU] numa: Add support for CPU-less HBM/NUMA node binding
- [Fix] scheduler: Publish KV metrics when idle
- [NPU] Enable NGRAM speculative decoding on Ascend NPUs
- [Feat] model: Add MiniCPM-OCR 1.0 support
- [Feat] observability: Log stats lines for idle steps on PD prefill and decode workers
- [Fix] speculative: Accept draft probabilities that round just above 1 in chain verify
- [Fix] dcp: Use persistent buffers for graph-captured a2a
- Docs
- Python not yet supported