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
- [diffusion] perf dump: one writer per replica, no first-request NCCL setup
- Serve decision model checkpoints natively on /v1/decisions, /v1/jev, and /v1/systemone
- [diffusion] Keep the DiT off component offload under explicit multi-GPU FSDP
- [diffusion] Fix the OOM hint's placement advice and add a 24 GB Wan2.1 recipe
- [Bug] `/v1/messages`: a `stop_sequences` hit is reported as `end_turn` with `stop_sequence: null`
- [Bug] `--enable-return-routed-experts` returns all-zero routings on the triton-kernels / flashinfer top-k paths
- [HiCache] Attribute buffer-mode storage hits against the joint device match
- [Perf] Coalesce the deduplicated conv-window store in causal_conv1d_update
- fix: abort parallel sampling requests by caller rid
- [Rust] Implement the Rust gRPC adapter for native generation
- Docs
- Python not yet supported