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] Check tensor dtypes in the E8M0 parameter copy guard
- [Fix][Humming] Honor input quantization ignore aliases
- [DP] consistent_hash load-balance method with optional load-aware spill + repin
- [fix] Mllama image collection and cross-attention masking
- [Fix] Make PLE logical clearing safe while KV memory is offloaded
- [Fix] Harden Apple Silicon MLX install against transformers 4.x drift (#38183)
- fix(symm-mem): disable multimem all-gather when the local GPUs have no peer access
- docs: add Halo post-training integration
- [unified-memory] Admit EAGLE/EAGLE3 onto fused draft KV, on mamba-MHA and MLA hosts
- [unified-memory] Admit DFLASH and DSPARK with fused draft KV
- Docs
- Python not yet supported