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] scheduler: add fixed-batch bounded LPM policy
- [Feat] mm: Add --disable-multimodal to serve a multimodal model text-only
- [Perf] disaggregation: Add the just_in_time decode allocation policy
- [Feat] [Intel] dwdp: Add a Level Zero VMM backend for DWDP on XPU
- [Fix] srt: Make DWDP, device timer, and AMX tests platform-safe
- [CI] [AMD] rerun-test: Add `--amd` to rerun tests on their AMD runners
- [Refactor] dwdp: Route DWDP through a device-neutral VMM backend
- [Feat] spec: add per-request speculative decoding opt-out with mixed-batch support (#47797)
- [Fix] models: Set k/v_scale_float when loading KV cache scales from file
- [Fix] dwdp: Retry local VMM allocation after CUDA OOM
- Docs
- Python not yet supported