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
- [RFC] Context-parallel support for the unified L3 key scheme
- [Fix] deterministic: Disable mixed chunked prefill
- [Fix] reasoning_parser: Promote a closed think block to content at stream end under force_nonempty_content
- [Feat] modelopt: Support packed GGML mixed-precision weights
- [Fix] disaggregation: Release multimodal features when PD prefill finishes a request
- [Fix] modelopt: Preserve explicitly quantized NextN drafts
- [Feat] multimodal: Register fetchers for custom media URL schemes
- [Fix] weight-export: handle missing send groups without KeyError
- [Perf] speculative: Cache the grammar mask host token filter and skip unconstrained draft-tree walks
- [AMD] Run MiniMax-M3 decode layers as mono kernels on gfx950
- Docs
- Python not yet supported