We will now start with our implementation of a convolutional neural net in Keras. For our example case, we will train a network to classify Fashion-MNIST. This is a dataset of grayscale images of fashion products, of the size 28 x 28. The total number of images is 70,000, with 60,000 as training and 10,000 as a test. There are ten categories in this dataset, which are t-shirt, trousers, pullover, dress, coat, sandal, shirt, sneakers, bag, and ankle boots. Labels for each are marked with a category number from 0-9.
We can load this dataset as follows:
from keras.datasets import fashion_mnist
(x_train, y_train), (x_test, y_test) = fashion_mnist.load_data()
The previous code block doesn't output a visualization of the dataset, so following image is to show what dataset we will be using:
It will split the data into the train and test sets with both inputs...
