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 supported43 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Bug]: vLLM crashes during generation with LMCacheMPConnector on Qwen3.6/Qwen3.5 hybrid model: AssertionError "External KV connector is not verified yet"
- fix: remove extraneous 'messages' argument in DeepseekV4Renderer chat template call (closes #44949)
- fix: correct operator precedence in use_existing_torch.py (closes #42363)
- [Bugfix] Abort requests that arrive while the engine is asleep
- Reduce prompt logprobs memory usage
- [Quant] Port RMSNormQuantFusionPass to manual fusion (fp8 static per-tensor + dynamic per-token)
- [Bug]: Performance Regression since 0.22 on RDNA2
- [Bugfix][Spec Decode] Fix drafter slot mapping type mismatch when DBO is enabled
- [Bug] Ray executor fails at EngineCore init with ActorHandleNotFoundError on ray 2.55.1 (works on CI-pinned 2.48.0)
- [Feature] Fused K-RoPE + static FP8 per-tensor KV Cache write
- Docs
- Python not yet supported