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 hybrid-linear KV transfer under prefill pipeline parallelism
- [Bug] EXIF orientation is never applied when decoding images for VLM serving
- [Bug] DSPARK retrieves wrong key "acc_linear_penalities" from sampling_info
- Title: Inconsistent error response format with OpenAI standard for non-streaming requests
- [Bug] The --json-model-override-args does not recursively update the model configuration.
- [Bug][PD] Prefill request aborted during KV transfer returns spurious finish_reason "length" (200) instead of "abort" (400/503)
- [bugfix] all reduce hidden_states from mlp in last decoder layer for Qwen3NextModel
- [docs] GLM-Image high-performance commands
- [Bug] Encountered an error while loading the Minimaxh3 model.
- [Bugfix][Sampler] Keep tied top-k tokens in the PyTorch sampling fallback
- Docs
- Python not yet supported