Most deep learning papers use a square image to train the network, but I was wondering if rectangular images would be any different.
The dateset I have currently cons
