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(metrics): refresh idle DP gauges after empty batches
- [LoRA] Bound expert_id against num_experts in fused MoE LoRA kernel
- [DSV4.1][PD][1/N] Support prefill node use cp parallel strategy
- fix: accept pp_proxy_tensors in frozen-kv-mtp, uno, and multi-layer-eagle forward_batch_generation
- [DSV4.1][PD][7/N][Perf] Reuse shared compressed KV dequantization in DeepSeek V4.1 CP prefill
- [DSV4.1][PD][8/N][Perf] Fuse candidate selection and masking for DeepSeek V4.1 prefill
- [Scheduler] Reclaim the idle device cache on every rank
- [DCP] Run fi_a2a on FlashInfer's fused all-to-all + LSE reduce
- [HiCache] Fail fast on a null pointer from the Mooncake host allocator
- [diffusion][AMD] FlyDSL norm: pick the wave count per hidden size
- Docs
- Python not yet supported