You have an Azure Machine Learning workspace and a deep learning model that requires a GPU. You need to deploy this model as a web service.

Which two compute targets can you use to host the GPU-enabled web service? Each correct answer presents a complete solution.