Hands-On Deep Learning Algorithms with Python

Generate the handwritten digit, 7

We set the digit (label) to generate as 7:

label_to_generate = 7onehot = np.eye(10)

Set the number of iterations:

for epoch in range(num_epochs):    for i in range(len(images) // batch_size):

Sample images based on the batch size:

        batch_image = images[i * batch_size:(i + 1) * batch_size]

Sample the condition that is, digit we want to generate:

        batch_c = labels[i * batch_size:(i + 1) * batch_size]

Sample noise:

        batch_noise = np.random.normal(0, 1, (batch_size, 100))

Train the generator and compute the generator loss:

        generator_loss, _ = session.run([D_loss, D_optimizer], {x: batch_image, c: batch_c, z: batch_noise})    

Train the discriminator and compute the discriminator loss:

 discriminator_loss, _ = session.run([G_loss, ...