Collapse by Conditioning: Training Class-conditional GANs with Limited Data

by   Mohamad Shahbazi, et al.
ETH Zurich

Class-conditioning offers a direct means of controlling a Generative Adversarial Network (GAN) based on a discrete input variable. While necessary in many applications, the additional information provided by the class labels could even be expected to benefit the training of the GAN itself. Contrary to this belief, we observe that class-conditioning causes mode collapse in limited data settings, where unconditional learning leads to satisfactory generative ability. Motivated by this observation, we propose a training strategy for conditional GANs (cGANs) that effectively prevents the observed mode-collapse by leveraging unconditional learning. Our training strategy starts with an unconditional GAN and gradually injects conditional information into the generator and the objective function. The proposed method for training cGANs with limited data results not only in stable training but also in generating high-quality images, thanks to the early-stage exploitation of the shared information across classes. We analyze the aforementioned mode collapse problem in comprehensive experiments on four datasets. Our approach demonstrates outstanding results compared with state-of-the-art methods and established baselines. The code is available at:


page 2

page 9

page 17

page 18

page 19

page 20

page 21


Virtual Conditional Generative Adversarial Networks

When trained on multimodal image datasets, normal Generative Adversarial...

Class-Splitting Generative Adversarial Networks

Generative Adversarial Networks (GANs) produce systematically better qua...

Efficient Conditional GAN Transfer with Knowledge Propagation across Classes

Generative adversarial networks (GANs) have shown impressive results in ...

Rebooting ACGAN: Auxiliary Classifier GANs with Stable Training

Conditional Generative Adversarial Networks (cGAN) generate realistic im...

Interpreting CNN Predictions using Conditional Generative Adversarial Networks

We propose a novel method that trains a conditional Generative Adversari...

Distilling Representations from GAN Generator via Squeeze and Span

In recent years, generative adversarial networks (GANs) have been an act...

Searching towards Class-Aware Generators for Conditional Generative Adversarial Networks

Conditional Generative Adversarial Networks (cGAN) were designed to gene...

Please sign up or login with your details

Forgot password? Click here to reset