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
- [Fix] Correct MiniCPM fused top-k synchronization and selection
- fix(dsv4): report dynamic FP8 KV cache scales
- fix(b12x): derive scratch capacity from scheduler contract
- [Bug] DeepseekV4ForCausalLM hook rejects `--speculative-algorithm NEXTN` because the NEXTN->EAGLE alias is resolved after model-specific hooks
- [NPU] Refactor partial is_npu call-sites
- fix(qwen3_5): make a bypassed-load weight update reach ParameterMapper and the derived buffers
- [HiSparse] Fix swap-in planner prefix-scan corruption below swap_in_block_size 512
- Fix #38236: accept NEXTN as an alias for EAGLE in DeepseekV4 hook
- [NPU] Enable short-conv hybrid models (LFM2/LFM2-MoE) with sgl_kernel…
- fix(scheduler): enable WAR barrier on HIP/ROCm to prevent overlap scheduler race
- Docs
- Python not yet supported