Built-in datasets in TF 2.0

TF 2.0 also provides a collection of datasets that are ready to be used with TensorFlow. It handles downloading, preparing the data, and even building tf.data.Dataset on its own, which can then be directly fed into the model.

Follow these steps to use these built-in datasets:

1. Install the TensorFlow datasets:

pip3 install tensorflow-datasets

Please note that tensorflow-datasets expects you to have a correct and complete installation of TF 2.0.

2. After tensorflow-datasets has been installed, you can view a list of available datasets by using the following code:

import tensorflow_datasets as tfds
tfds.list_builders()

This will give the following output:

['abstract_reasoning', 'bair_robot_pushing_small', 'caltech101', 'cats_vs_dogs', 'celeb_a', 'celeb_a_hq', 'chexpert', 'cifar10...