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 supported51 Subscribers
View all SubscribersAdd a CodeTriage badge to vllm
Help out
- Issues
- [Feature]: Re-enable `TRITON_UNFUSED` MXFP4 MOE backend
- [Perf][Model Loader] Reload layout-identical weights directly
- [Bug]: StructuredOutputManager sizes grammar-compile ThreadPoolExecutor by host CPU count — cgroup-unaware and uncapped, causing CFS throttling and decode stalls on Kubernetes
- [Bugfix][Structured Output] Cap grammar-compile executor workers to avoid CFS throttling in containers
- fix: accept empty tool call list in case of none required tool_choice
- [Bug]: `MooncakeStoreConnector` KV Offload hangs the vLLM engine on a GDN model (`Qwen3.6-35B-A3B`)
- [Bugfix] Fix DBRX weight loading crash after RoutedExperts refactor (#41184)
- [Usage]: DeepSeek-V4-Flash on single B300 — auto-enabled VLLM_USE_BREAKABLE_CUDAGRAPH caps throughput (disabling gives ~1.6x); is it safe, and is torch.compile support planned?
- [XPU][DSv4] Use SYCL deepseek_inv_rope_fp8_quant kernel in DeepSeek-V4 o_proj
- [RFC] [Feature] [Experimental] Support GigaToken Accelerated Tokenizer Mode
- Docs
- Python not yet supported