Convert cifar-10 dataset to images dataset for using tf.keras.preprocessing.image.ImageDataGenerator.flow_from_directory
Download the data here : https://www.cs.toronto.edu/~kriz/cifar.html
Extract the zip file and move it to the folder.
Create the cifar-10 folder.
Run the script.py, you can modify the rate between validation set and testing set.
Look at these 2 images for more better intuition.
Cifar-100 is similar, modify the classes list and the number of classes.
Instant data : https://www.kaggle.com/hngbiquc/cifar10-images-dataset
Your folder look like this before running the script. Your folder look like this after running the script.