MeuXina: Accelerating Generative Adversarial Networks with HPC
Generative Adversarial Networks (GAN) are a type of Deep Neural Network (DNN) containing at least two networks within their structure – a Generator and a Discriminator. The Generator and the Discriminator play a competitive game against each other: The Generator tries to synthesize the most realistic data, e.g., images, videos, or signals, while the Discriminator is trained to distinguish real from fake (generated) data. The aim is to achieve equilibrium between these two neural networks, which requires a huge amount of data as well as a large amount of computation.
