We describe a new training methodology for generative adversarial networks. The key idea is to grow both the generator and discriminator progressively: starting from a low resolution, we add new ...

After hours: January 10 at 7:44:42 PM EST ...