Imagenet example with inception v3

vision
11

Ok, so you just do loss1 + loss2 like here How to extract features of an image from a trained model ? Then backprop them?

1 Like