A senior machine learning engineer is looking to deploy a model as a Docker container and expose it as a REST API in a production environment using MLflow. Which MLflow command or functionality should they use to create a Docker image for the model?