Skip to content

Instantly share code, notes, and snippets.

@vivekpadia70

vivekpadia70/dcgan_mnist.py

Last active March 30, 2020 10:26
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save vivekpadia70/46bacc09a4e9e4e67243fe8fcd7e0e4c to your computer and use it in GitHub Desktop.
Save vivekpadia70/46bacc09a4e9e4e67243fe8fcd7e0e4c to your computer and use it in GitHub Desktop.
Download ZIP
Raw
dcgan_mnist.py
cross_entropy = tf.keras.losses.BinaryCrossentropy(from_logits=True)
def D_loss(real_output, fake_output):
real_loss = cross_entropy(tf.ones_like(real_output), real_output)
fake_loss = cross_entropy(tf.zeros_like(fake_output), fake_output)
total_loss = real_loss + fake_loss
return total_loss
def G_loss(fake_output):
return cross_entropy(tf.ones_like(fake_output), fake_output)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment