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 supported59 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- support qwen3.8-flash-next pp
- [Bug]: MRV2 extensible KV cache commits all KV before encoder CUDA graph capture, so cudagraph_mm_encoder OOMs at startup
- [XPU][CI] fix the wrong test name in intel CI
- [Bug]: Nemotron-3.5-Super-VL-120B-A12B-NVFP4 native MTP fails to load: "start (0) + length (5376) exceeds dimension size (2688)"
- [KV Offload] Prefetch secondary lookups on request admission
- [Perf][Model Runner V2] Skip no-op staged-write flush of BlockTables.num_blocks
- [CI] Let ci-selector decide CMakeLists.txt changes on CUDA
- [Perf] Cache per-call host work in MXFP8 linear, NVFP4 KV views, NVFP4 MoE and router GEMM
- fix(mamba): stop a mid-block resume at the block boundary
- [HiSparse] Alternative to #55398: promote P/D imports into GPU pages without NIXL changes
- Docs
- Python not yet supported