The training loop remains unchanged, with the exception of the discriminator build call, which requires new dimensions that correspond to CIFAR-10's image size, as shown in the following code:
discriminator = build_discriminator(img_shape=(32, 32, 3))
It would often be the case that we would need to adapt our learning rates, or the network architecture, when moving from one dataset to the other; luckily, that's not the case in this example.