DCGAN is initialized with random weights, so a random code plugged into your network would crank out a totally random image. However, when you might imagine, the network has millions of parameters that we will tweak, and also the purpose is to locate a environment of those parameters which makes samples generated from random codes appear like the s