Imagenet example with inception v3

vision
4

you need to compute loss on both x and aux (separately). aux is the auxillary classifier referred to in the inception paper.

1 Like