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
- [Bug] sgl-deep-gemm 0.2.0: SM90 weight-scale transform returns a non-owning alias (follow-up to #32188)
- Add zero-copy stacked-embedding CUDA IPC transport for the score API
- [Bug] Hierarchical cache returns wrong output on a hybrid (GDN/Mamba) model: a full-length host-tier hit answers in a different pass's voice, 20/20, on main
- [RFC] Align KV cache events with vLLM's schema so shared consumers have a standard format
- [DSpark][NPU] Fix attn_tp > 1 draft wo_b all_reduce group
- Add checkpoint-driven response-template adapters
- [Bug] Illegal memory access in the Triton fused-MoE kernel when `flashinfer_megamoe` is combined with EAGLE speculative decoding (GLM-5.2-NVFP4, sm_107)
- [RL] Add /pull_weights: engine-side pull of published weights into a host-local checkpoint
- [Fix] Resolve Inkling think-end IDs in ReasonerGrammarBackend and gate the canonical-schema bypass
- [bugfix] chunked prefill req counted as a new waiting req in calculating batch_is_full
- Docs
- Python not yet supported