Are topographic deep convolutional neural networks better models of the ventral visual stream?