Working with TensorFlow Estimators

TensorFlow estimators are also reusable components. The Estimators are higher-level APIs that enable users to build, train, and deploy machine learning models. It has several pre-made models that can save users from the hassle of creating computational graphs or sessions. This makes it easier for users to try different model architectures quickly with limited code changes. The Estimators are not specifically dedicated to deep learning models in the same way as tf.keras. Therefore, you will not find a lot of pre-made deep learning models available. If you need to work with deep learning frameworks, then the tf.keras API is the right choice to get started.

For this example, we are going to set up the same regression problem and build a regression model. The source of data is the same one we used in streaming training data, which is available through Google Cloud's BigQuery:

DATASET_GCP_PROJECT_ID = 'bigquery-public-data'
DATASET_ID...