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
- [ROCm] aiter backend: write extend attention output into the PCG-provided buffer
- [ROCm] extend-attention: window-aware long-extend tile on gfx950 (bit-identical, −5% TTFT at 64k uncached)
- [ROCm] rope_cache: group fused qk-rope+cache programs per KV group (bitwise exact, 2.15-2.5x on gfx950)
- [Quant] Preserve ModelOpt FP4 quantization for NextN drafts
- [Bugfix] Preserve parent swa_uuid when compacting SWA radix chain
- [CP] Fix CP-v2 all-gather not using symmetric memory
- fix(observability): restore speculative-decoding spans on the Spec V2 paths
- [MLX] Fall back to no_buffer mamba radix cache on Apple Silicon
- [diffusion] feat: expert parallelism for MoE DiTs
- [CPU] Part 1: Add DeepSeek3.x/GLM5.x optimized sgl-kernel for CPU
- Docs
- Python not yet supported