Tensorizing Generative Adversarial Nets
This work addresses the deployment of GANs on platforms with limited computational power, such as mobile phones, by providing a method to compress models without sacrificing quality.
The paper tackles the problem of high computational complexity and large parameter counts in Generative Adversarial Networks (GANs) by introducing a tensor-based framework that reduces model parameters by up to 35 times on the MNIST dataset while maintaining generative performance.
Generative Adversarial Network (GAN) and its variants exhibit state-of-the-art performance in the class of generative models. To capture higher-dimensional distributions, the common learning procedure requires high computational complexity and a large number of parameters. The problem of employing such massive framework arises when deploying it on a platform with limited computational power such as mobile phones. In this paper, we present a new generative adversarial framework by representing each layer as a tensor structure connected by multilinear operations, aiming to reduce the number of model parameters by a large factor while preserving the generative performance and sample quality. To learn the model, we employ an efficient algorithm which alternatively optimizes both discriminator and generator. Experimental outcomes demonstrate that our model can achieve high compression rate for model parameters up to $35$ times when compared to the original GAN for MNIST dataset.