The Inception V3 network is most often seen using images that are 299x299 as input. I am trying it now on images of 599x599 -- its taking a long time to train, so I have ti
