TensorFlow 2.0 RC is available Learn more

Module: tf.contrib.gan.losses

View source on GitHub

TFGAN losses and penalties.

Losses can be used with individual arguments or with GANModel tuples.

Modules

wargs module: TFGAN grouped API. Please see README.md for details and usage.

Functions

acgan_discriminator_loss(...)

acgan_generator_loss(...)

combine_adversarial_loss(...): Combine adversarial loss and main loss.

cycle_consistency_loss(...): Defines the cycle consistency loss.

least_squares_discriminator_loss(...)

least_squares_generator_loss(...)

minimax_discriminator_loss(...)

minimax_generator_loss(...)

modified_discriminator_loss(...)

modified_generator_loss(...)

mutual_information_penalty(...)

stargan_discriminator_loss_wrapper(...): Convert a discriminator loss function to take a StarGANModel.

stargan_generator_loss_wrapper(...): Convert a generator loss function to take a StarGANModel.

stargan_gradient_penalty_wrapper(...): Convert a gradient penalty function to take a StarGANModel.

wasserstein_discriminator_loss(...)

wasserstein_generator_loss(...)

wasserstein_gradient_penalty(...)