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
- Docs: official BFL FLUX.2-klein FP8 is Comfy split-QKV, not a packed SGLang --transformer-weights-path drop-in
- Diffusion CUDA JIT: rsqrt host/CRT clash burns tens of seconds before soft-fail (fail fast / skip on sm<80)
- [AMD][DSv4] Run selected DSv4 attention projections in MXFP4
- fix(gateway): replace, don't accumulate, when a worker URL is re-registered
- fix(gateway): give an in-flight AddWorker a head start before removing the last worker
- Fix FlashInfer planning for configured attention logits soft cap
- [NPU] Add Kimi-K3 W4A8 2P4D PD-separation accuracy test (gpqa)
- [test] Fix image prompts in send_one.py
- [Fix] Hold back the longest partial tool-call marker when streaming
- [Fix] Close a Harmony final block at <|end|>, not only <|return|>
- Docs
- Python not yet supported