DigGAN: Discriminator gradIent Gap Regularization for GAN Training with Limited Data
Preprint
Abstract
Generative adversarial nets (GANs) have been remarkably successful at learning to sample from distributions specified by a given dataset, particularly if the given dataset is reasonably large compared to its dimensionality. However, given limited data, classical GANs have struggled, and strategies like output-regularization, data-augmentation, use of pre-trained models and pruning have been shown to lead to improvements. Notably, the applicability of these strategies is 1) often constrained to particular settings, e.g., availability of a pretrained GAN; or 2) increases training time, e.g., when using pruning. In contrast, we propose a Discriminator gradIent Gap regularized GAN (DigGAN) formulation which can be added to any existing GAN. DigGAN augments existing GANs by encouraging to narrow the gap between the norm of the gradient of a discriminator's prediction w.r.t.\ real images and w.r.t.\ the generated samples. We observe this formulation to avoid bad attractors within the GAN loss landscape, and we find DigGAN to significantly improve the results of GAN training when limited data is available. Code is available at \url{https://github.com/AilsaF/DigGAN}.Keywords
All Related Versions
This publication has 10 references indexed in Scilit:
- Ensembling Off-the-shelf Models for GAN Training2022
- Few-shot Image Generation via Cross-domain Correspondence2021
- ReMix: Towards Image-to-Image Translation with Limited Data2021
- PacGAN: The Power of Two Samples in Generative Adversarial NetworksIEEE Journal on Selected Areas in Information Theory, 2020
- Self-Ensembling With GAN-Based Data Augmentation for Domain Adaptation in Semantic Segmentation2019
- Sliced Wasserstein Generative Models2019
- Detecting Overfitting of Deep Generative Networks via Latent Recovery2019
- Max-Sliced Wasserstein Distance and Its Use for GANs2019
- On the Effectiveness of Least Squares Generative Adversarial NetworksIEEE Transactions on Pattern Analysis and Machine Intelligence, 2018
- Generate to Adapt: Aligning Domains Using Generative Adversarial Networks2018