Spectral Normalization for Generative Adversarial Networks
This addresses a key training challenge in GANs for image generation, offering a computationally light and easy-to-implement solution.
The paper tackles the instability in training generative adversarial networks by proposing spectral normalization for the discriminator, resulting in SN-GANs that generate images of better or equal quality on datasets like CIFAR10 and ILSVRC2012.
One of the challenges in the study of generative adversarial networks is the instability of its training. In this paper, we propose a novel weight normalization technique called spectral normalization to stabilize the training of the discriminator. Our new normalization technique is computationally light and easy to incorporate into existing implementations. We tested the efficacy of spectral normalization on CIFAR10, STL-10, and ILSVRC2012 dataset, and we experimentally confirmed that spectrally normalized GANs (SN-GANs) is capable of generating images of better or equal quality relative to the previous training stabilization techniques.