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] cuda-graph: Raise the decode graph ceiling default to 128 for 35 to 60 GiB GPUs
- [Fix] frontend: Preserve zero logprobs in choice normalization
- [Fix] disaggregation: bound staging requests by destination pages
- [Perf] dsv41: Run the decoder tail only over the prompt's last window
- [Fix] diffusion/cli: Preserve explicit argv and registered aliases
- [Fix] rust-server: Align Jinja chat inputs with Python
- [Perf] batch-invariant: Fused residual add + RMSNorm kernel
- [Perf] http: Serialize /v1/embeddings responses with model_json_response
- [CI] rust: Publish sglang-api-types on crates.io
- [Fix] jit: Fix permanent cache misses behind a symlinked cache dir
- Docs
- Python not yet supported