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] FP4 KV block scales are dropped by PD transfer and cpu_tensor retraction (fp4_mx_block16 gives garbage)
- [Fix] multimodal: Return the selected processor from LlavaMultimodalProcessor
- [CI] [AMD] diffusion: Set the retry deadline on the multimodal lanes
- [CI] [AMD] diffusion: Run more than one multimodal-gen matrix job at a time
- [Perf] [NVIDIA] attention: Native FP8 KV cache for SM120 FA4 kernels
- feat(diffusion): add opt-in plane-major Ulysses QKV layout
- Hsr/minimax h3 subblock sage fp8
- [Perf] DFLASH: route draft bf16 GEMMs through tuned dispatch on GB300
- [Feat] nemotron3-embed: Support Nemotron-3-Embed embedding models
- [Fix] InstantTensor: return load-time GPU memory to the KV cache
- Docs
- Python not yet supported