Skip to content
Snippets Groups Projects
Commit 1f0df5ff authored by Jean-Luc Parouty's avatar Jean-Luc Parouty
Browse files

Update DCGAN

parent 30323274
No related branches found
No related tags found
No related merge requests found
source diff could not be displayed: it is too large. Options to address this: view the blob.
......@@ -71,6 +71,9 @@ class DCGAN(keras.Model):
loss_function : Loss function
'''
super(DCGAN, self).compile()
self.discriminator.compile(optimizer=discriminator_optimizer, loss=loss_function)
self.generator.compile(optimizer=generator_optimizer, loss=loss_function)
self.d_optimizer = discriminator_optimizer
self.g_optimizer = generator_optimizer
self.loss_fn = loss_function
......@@ -121,11 +124,11 @@ class DCGAN(keras.Model):
combined_images = tf.concat( [generated_images, real_images], axis=0)
# Creation of labels corresponding to real or fake images
# 1 is generated, 0 is real
labels = tf.concat( [tf.ones((batch_size, 1)), tf.zeros((batch_size, 1))], axis=0)
# 0 is generated, 1 is real
labels = tf.concat( [tf.zeros((batch_size, 1)), tf.ones((batch_size, 1))], axis=0)
# Add random noise to the labels - important trick !
labels += 0.05 * tf.random.uniform(tf.shape(labels))
# labels += 0.05 * tf.random.uniform(tf.shape(labels))
# ---- Train the discriminator -----------------------------
# ----------------------------------------------------------
......@@ -155,7 +158,7 @@ class DCGAN(keras.Model):
random_latent_vectors = tf.random.normal(shape=(batch_size, self.latent_dim))
# Assemble labels that say all images are real, yes it's a lie ;-)
misleading_labels = tf.zeros((batch_size, 1))
misleading_labels = tf.ones((batch_size, 1))
# ---- Train the generator ---------------------------------
# ----------------------------------------------------------
......
0% Loading or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment