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
- vllm sglang hot reload
- [Bug] Pretokenized image requests can crash GLM MRoPE with a stale retokenized mask
- perf(spec): add vocab-parallel MTP local argmax
- [2/4][Kernels] Add MXFP4 KV cache codec (Triton quantize/dequant)
- [3/4][Kernel] Fused MXFP4 DSV4 decode attention for SM90 (FlashMLA-style, JIT)
- [4/4][DeepSeek V4] MXFP4 KV cache for the DSV4 serving stack on Hopper
- [Perf] Lossless final layer prefill optimization for LLaMA via Single-Query Attention
- Fix HiCache backup writer selection with context parallelism
- [Feature] Tuning / per-GPU config for DSv4 top-k v2 cluster launch plan (kClusterFloor / kNumPersistentClusters / kCandidates)
- LMCache integration ignores per-request cache_salt: external storage keys are token-only, breaking tenant isolation
- Docs
- Python not yet supported