Cloud ML is a managed service that allows you to train and host ML models without worrying about the underlying infrastructure. It provisions all of the requisite resources.
You can accelerate the learning process since a range of CPU, GPU, and TPU nodes are supported. It works with multiple frameworks, but the most popular is TensorFlow. As TensorFlow is open source, it allows for portability. Models can be trained locally on limited data and then sent to GCP to train at scale. Cloud ML integrates with other GCP services, such as Cloud Storage for data storage and Cloud Dataflow for data processing.
