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
- [Perf] pp: Defer steady decode output packing until after forward submission
- [CI] diffusion: Allow 6% host anon memory headroom on 5090
- [Refactor] mem_cache: Tag full-only KV frees that release a tree node's value, and commit load-backs through one method
- [CI] Isolate editable install cache locks per runner
- [Perf] hybrid: Reuse cached full-attention KV during GDN state replay
- [Test] disaggregation: Give the fake draft pool follows_target_swa_ids
- [Test] diffusion: Pass --trust-remote-code for the custom VAE path case
- [Fix] [NPU] eplb: preserve expert storage and reuse TP communication
- [Fix] openai: Accept `reasoning` as an alias for assistant `reasoning_content`
- [Fix] [AMD] minimax: Run the gfx950 Gluon sparse prefill with several KV heads per rank
- Docs
- Python not yet supported