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
- [Feat] scheduler: add mixed-prefill resource controls
- [Fix] ipc: Make multimodal payloads, hidden states and weight-update dtypes round-trip under msgpack IPC
- [Feat] scheduler: Add max-queued-tokens for request management
- [Fix] deepseek_v41: track image work after cancellation
- [Fix] benchmark: Add offline guidance for random datasets
- [Fix] scheduler: Do not request input logprobs again when re-prefilling a retracted request
- [Fix] rust-server: End an /abort_request-ed request on the scheduler's abort finish reason, as Python does
- [Perf] [NVIDIA] glm5: Route small-M BF16 GEMMs to FlashInfer warp split-K with PDL on SM107
- [Feat] diffusion: Support Pika-Labs/Pika-H3-Flex (MiniMax-H3 + 6-step LoRA)
- [Feat] mistral3: Support Nemotron 3.5 Rerank multimodal reranking
- Docs
- Python not yet supported