There is currently no official vLLM Docker image for Mac (Apple Silicon) CPU on Docker Hub. The official vLLM images on Docker Hub (vllm/vllm-openai) are built for Linux/amd64 and require NVIDIA GPUs; they do not support Apple Silicon CPUs out of the box. For Apple Silicon (M1/M2/M3), the official documentation and maintainers state you must build vLLM from source on your Mac, as there are no pre-built wheels or images for this device (docs, github_file).
There are some unofficial, community-built images (e.g., on Docker Hub by individual users), but there is no formal, supported vLLM CPU image for Mac M2Pro. The only official CPU image is published to AWS ECR (public.ecr.aws/q9t5s3a7/vllm-cpu-release-repo:v0.7.2), but this is for Linux/amd64, not Mac ARM (github_issue). If you want to run vLLM on your Mac M2Pro, you must upgrade your OS to meet the build requirements and build from source, or attempt to build a Docker image yourself using the provided Dockerfiles and cross-compilation instructions (docs, github_file).
Would you like more detail on how to build vLLM from source or create a custom Docker image for Apple Silicon?
Sources: