vllm
https://github.com/vllm-project/vllm
Python
A high-throughput and memory-efficient inference and serving engine for LLMs
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 supported26 Subscribers
Add a CodeTriage badge to vllm
Help out
- Issues
- Add padding to support TP with Marlin linear kernel (NVFP4 on H100)
- [Multimodal] Simplify ViT CUDA graph interfaces
- Enable B12x backend for non-gated MoEs (like Nemotron)
- [DP] External DP LB compatibility with all GPU visibility
- [Bugfix][Spec Decode] Properly share dpsk-v4 topk_indices_buffer.
- [Bugfix] Use enable_sm120_family for per-tensor FP8 CUTLASS kernels on SM12.1
- [Compilation] Port QKNormRoPEFusionPass and AsyncTPPass to VllmFusionPatternMatcherPass
- Enable DBO with elastic EP on NIXL EP
- [Bugfix][Hybrid][NemotronH] Fix mamba_cache_mode=all + speculative decoding crash
- Abort removed-rank requests during Elastic EP scale-down
- Docs
- Python not yet supported