Converting BigQuery ML models into TensorFlow

In this section, we'll train the same deep neural network that we trained in Chapter 11, Implementing Deep Neural Networks, and then export this model into the TensorFlow SavedModel format.

Training the BigQuery ML to export it

Before we start training the model, let's access BigQuery to create the dataset and the tables that will be used for training and prediction:

1. Log into our Google Cloud Console and access the BigQuery user interface from the navigation menu.
2. Create a new dataset under the project that we' created in Chapter 2, Setting Up Your GCP and BigQuery Environment. For this use case, we'll create a dataset called 13_tensorflow_model with the default options.
3. Now, we're ready to create the table that will contain the training dataset. Let's execute the following SQL statement:

   CREATE OR REPLACE TABLE `13_tensorflow_model.training_table` AS
         ...