DCGAN is initialized with random weights, so a random code plugged to the network would produce a completely random picture. However, when you might imagine, the network has millions of parameters that we can tweak, and the intention is to locate a environment of such parameters which makes samples generated from random codes look like the schoolin