Questions about loss function

[XiajieWei]

Hi, I'm confused about the loss function descripted in the paper and the corresponding implementation. I don't understand the implementation of L_dis in the code. In VAE L_pixel corresponds to MSE, can I just replace the image pixel data with discriminator hidden layer output Dis_l(x), if not, can you give more details in your implementation?