Got it, sounds like you would like to be able to reload a new model as the API Server is running in production?