I am following this tutorial about creating a generative variational auto-encoder. However, I am using images with 3 channels (RGB) of dogs pulled from Kaggle. I am building my