DCGAN is initialized with random weights, so a random code plugged into the network would deliver a totally random graphic. Even so, as you might imagine, the network has countless parameters that we can easily tweak, along with the intention is to find a location of these parameters that makes samples created from random codes look like the teachi