DCGAN is initialized with random weights, so a random code plugged to the network would make a totally random picture. Even so, while you may think, the network has many parameters that we can tweak, and the goal is to locate a placing of those parameters that makes samples generated from random codes look like the training data.We’ll be getting