Input and decoder output value ranges

[royhj]

Thank you for sharing this repo! In VanillaVAE (and maybe others) I was wondering about the choice of using tanh as the final activation for outputs (that has a range of [-1,1]) without normalizing the input images to the same range [-1,1] (in the dataset transforms). Wouldn't it make the reconstruction loss work much harder, opposed to either normalizing the input or using a final activation like sigmoid?

[HareshKarnan]

I had the same question. I think the output should be sigmoid and not tanh.