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
- [Multimodal] Generalize request-local CUDA IPC feature packing
- [Feature] Support EAGLE3 chain decoding for MiniCPM-SALA
- [Feature] Support DSpark decoding for MiniCPM-SALA
- [NPU][CI] Add CANN 9.1.0 nightly suites and mirror 6 tests onto the 9.1.0 image
- [Bugfix] MoE runtime LoRA: nondeterministic logprobs and greedy flips on identical inputs (temperature=0, small-batch decode)
- Stop --default-chat-template-kwargs from outranking a request's reasoning_effort
- [decode] Spec-decode with large draft-token counts OOMs mid-traffic on the decode instance - no upfront capacity check (verify workspace vs KV pool is zero-sum)
- [Qwen3-Next] Fix cycling tokens in response for Qwen3-Next-80B-A3B
- Prefill context parallelism for Kimi K3
- fix(hicache): fail-soft Mooncake direct linker loads (resolve transfers in revalidate_load)
- Docs
- Python not yet supported