Your Discriminator takes only one image as input but in pix2pix paper it has two inputs one is input_image and generated image.