DCGAN is initialized with random weights, so a random code plugged into the network would make a completely random picture. However, while you may think, the network has an incredible number of parameters that we can tweak, along with the goal is to find a location of these parameters that makes samples created from random codes look like the educa