Deploying and predicting an ML model on GCP Cloud

Public cloud providers are offering several AI platforms for training built-in models, as well as your custom models, for deploying the models for predictions. Google offers the Vertex AI platform for ML use cases, whereas Amazon and Azure offer the Amazon SageMaker and Azure ML services, respectively. We selected Google because we assume you have set up an account with GCP and that you are already familiar with the core concepts of GCP. GCP offers its AI Platform, which is part of the Vertex AI Platform, for training and deploying your ML models at scale. The GCP AI Platform supports libraries such as scikit-learn, TensorFlow, and XGBoost. In this section, we will explore how to deploy our already trained model on GCP and then predict the outcome based on that model.

Google AI Platform offers its prediction server (compute node) either through a global endpoint (ml.googleapis.com) or through a regional endpoint (<region>-ml...