Getting error with keras custom loss function: __array__() takes 1 positional argument but 2 were given

Question

I am generating fake images using a pix2pix conditional Generative Adversarial Network (GAN). My GAN is working properly when I am using the mean_absolute_error (L1-distance) to