Here are some of Keras’s datasets (from the module tensorflow.keras.datasets¹³) for practicing deep learning. We’ll use these in the chapter’s examples, exercises and projects:

• MNIST¹⁴database of handwritten digits—Used for classifying handwritten digit images, this dataset contains 28-by-28 grayscale digit images labeled as 0 through 9 with 60,000 images for training and 10,000 for testing. We use this dataset in Section 16.6, where we study convolutional neural networks.

• Fashion-MNIST¹⁵ database of fashion articles—Used for classifying clothing images, this dataset contains 28-by-28 grayscale images of clothing labeled in 10 categories¹⁶ with 60,000 for training and 10,000 for testing. Once you build a model for use with MNIST, you’ll be able to reuse that model with Fashion-MNIST by changing a few statements. You’ll use this dataset in the exercises.

• IMDb Movie reviews¹⁷—Used for sentiment analysis, this dataset contains reviews labeled as positive (1) or negative (0) sentiment with 25,000 reviews for training and 25,000 for testing. We use this dataset in Section 16.9, where we study recurrent neural networks.

• CIFAR10¹⁸ small image classification—Used for small-image classification, this dataset contains 32-by-32 color images labeled in 10 categories with 50,000 images for training and 10,000 for testing. You’ll analyze this dataset with a convnet in the exercises.

• CIFAR100¹⁹small image classification—Also, used for small-image classification, this dataset contains 32-by-32 color images labeled in 100 categories with 50,000 images for training and 10,000 for testing. If you do the CIFAR10 exercise, you should be able to tweak your convnet model quickly for use with CIFAR100.