You are managing an Azure Machine Learning workspace that contains a batch endpoint. You plan to deploy a model to this batch endpoint. Which compute type should you use to configure the deployment?