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
- [OpenAI] Check each tool schema once, not on every request
- fix(tokenizer): two trust_remote_code tokenizer compat shims (padding_side, duplicate added-token ids)
- MultiDetokenizerRouter splits each batch into per-request IPC sends
- [PD] Preserve weight-version attribution across rebootstrap replay
- [Diffusion][NPU] Support MXFP8 attention for MiniMax H3 and relax sequence length constraints
- [Bug] expert-distribution endpoints terminate the scheduler when expert_distribution_recorder_mode is unset
- [Diffusion][NPU] Refactoring of sparse attention backends and support for A5
- [NPU] add int8 LightingIndexer option to speedup it by 2X on A3 and A2
- Fix initial reasoning-opener state after prefilled thinking
- [AMD] Small-M FP8 block-scale GEMM with fused activation quant for gfx950 (Qwen3.5)
- Docs
- Python not yet supported