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
- [Fix] Write the req_to_token page tail so rows stay valid over whole pages
- Fix multimodal draft embedding alignment
- [model-gateway] Add opt-in `fips` feature using the AWS-LC FIPS module
- [diffusion] Z-Image: autotune the shared transformer block at both token scales (H100 1024^2 denoise 13.58 -> 9.58 s, 1.42x)
- [dLLM] Preserve literal mask text in prompts
- [NVIDIA] perf: pipeline the extend-attention KV loads on Hopper
- Support Dream-v0-Base-7B
- [short-conv] Radix mamba-cache track for multi-entry conv state (1/3)
- [Model] Support DeepSeek-V4-Flash-Vision multimodal
- scheduler: opt-in DP-attention phase lockstep (--enable-dp-phase-lockstep)
- Docs
- Python not yet supported