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
- [Inkling] Fix inkling gate renorm nan
- [Perf][mem_cache] Persistent lazy eviction heap for the UnifiedRadixCache Full component (bit-identical order)
- [NPU] Integrate Kimi-K3 DSPARK, MXFP4 MoE and distributed inference support
- [Bug] DeepSeek-V4-Flash-Vision preview image: multi-turn tool calls come back wrapped in {"arguments": {...}}
- [Qwen4] Release weight-loader references without waiting for cyclic GC
- [Qwen4] Skip sampler token synchronization for single-rank groups
- [Qwen4] Preserve request cancellation after tokenizer cleanup
- [NPU]Fix timeout issues caused by network instability in multimodal test cases.
- supports_fp8() always True; _apply_fallback_scaled_mm still calls torch._scaled_mm (no escape on SM<89)
- sgl_kernel AOT: rmsnorm NoKernelImage on sm_75 while silu_and_mul / gelu_and_mul work (per-op cubin gap)
- Docs
- Python not yet supported