Experiment of using Instance Normalization vs Layer Normalization on the Decoder (MUNIT)

[tom99763]

Here are the results of using different normalization on the decoder.

By the computation operation of the normalization methods, the MUNIT architecture can be summarized as follows.

This means that since there's no tuning channel correlation on the upsampling layer (i.e., Adaptive Instance Normalization, StyleGAN), if you use instance normalization during upsampling, the tunned channel correlation (ResNet + Adaptive Instance Normalization) will be destroyed.