How to Create a Generative Adversarial Model

The way that generative models are created is through something called generative adversarial models. The first part generates the image and then the second part is discriminative where it tries to tell okay is this image generated by our model or is it like one of the real images that we see in the data. And essentially the discriminator is grading your homework and the better the discriminator does well, the more accurate its results will be.