DCGAN is initialized with random weights, so a random code plugged into the network would produce a totally random picture. On the other hand, when you may think, the network has millions of parameters that we can tweak, plus the objective is to find a placing of such parameters https://embedded-solutions53075.bloggip.com/34119303/how-much-you-need-to-expect-you-ll-pay-for-a-good-artificial-intelligence-platform