I am trying to understand how GANs work, while doing so, I got stuck at how the generator network is able to generate images from the random noise input, can someone explain