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 supported60 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Bugfix][MoE] Preserve invalid expert IDs in NIXL round-robin mapping
- [Bugfix][Kernel] Fix reshape & cache dropping KV heads past 512 chunks
- [Core][Spec Decode] Back only the tail of DFlash/DSpark sliding-window drafter KV
- [Bugfix][KV Offload]Defer async loads in mixed steps
- [ROCm][MLA][DCP] Serve the non-causal DSpark draft on TRITON_MLA under DCP
- [Bugfix][MRV2] Release the allocator cache between adaptive verification profile shapes
- [KV Offload] Add a store trigger and isolated store trigger handlers
- [ROCm] Route AITER Triton MXFP4 MoE over local experts under EP
- [ROCm][Kernel] Support persistent_topk on ROCm
- [KV Connector][NIXL] Report completed pulls before the next model output
- Docs
- Python not yet supported