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
- [Bug][kimi-k3] Long-context [PAD] (id 163839) storms + DSPARK inf/nan asserts — NaN-contaminated logits; released kimi-k3 image predates #32477; pad not stopped/filtered; allowed_special="all" makes [PAD] injectable
- [DeepSeek-V4] Support EAGLE3 for DeepSeek-V4-Flash-0731
- [Feature] Add opt-in bounded-load routing-key affinity to SGLang Model Gateway
- [Bugfix] Serialize disk weight-update completion state
- [Feature] Support dense NVFP4 W4A16 (bf16 activations) GEMM on SM120
- unified_cache: back up long prompts past their first (chunked) extend — currently skipped when extend exceeds chunked_prefill_size
- [Bugfix] Fix multimodal state in streaming sessions
- DSPARK enable for XPU
- fix: Allow remote-loaded instances to serve as TransferEngine weight seeds
- [speculative] DSpark: support the mask-filling draft convention (verify width = gamma)
- Docs
- Python not yet supported