Hands-On Deep Learning Algorithms with Python

Discriminator loss

The discriminator loss is given as follows:

As the discriminator loss of an InfoGAN is same as with a CGAN, implementing the discriminator loss is the same as what we learned in the CGAN section:

#real lossD_loss_real = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits=D_logits_real,  labels=tf.ones(dtype=tf.float32, shape=[batch_size, 1])))#fake lossD_loss_fake = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits=D_logits_fake,  labels=tf.zeros(dtype=tf.float32, shape=[batch_size, 1])))#final discriminator lossD_loss = D_loss_real + D_loss_fake