thanks for this cool work! I am a bit confused about table 2 , how do you evaluate exactly the "Accuracy of other
classes"? Do you generate 1 image per class of the 1000 classes of ImageNet or only of the classes in the table and then classify with ResNet?
Hi,
thanks for this cool work! I am a bit confused about table 2 , how do you evaluate exactly the "Accuracy of other classes"? Do you generate 1 image per class of the 1000 classes of ImageNet or only of the classes in the table and then classify with ResNet?
Thanks!