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 ...

I devoured an exceptional number of special meals over the past five months as I circled the Philadelphia region, fork in ...