Let's say the model was trained to translate images between cat, dog, tiger and lions. Given some pig images, the model can translate a cat image as content image to pig.
I wonder if it's possible to do something opposite. Given some pig images that the model never sees, is there a way to translate pig to cat?
In short, is there a way to translate unseen class to seen class?
Let's say the model was trained to translate images between cat, dog, tiger and lions. Given some pig images, the model can translate a cat image as content image to pig.
I wonder if it's possible to do something opposite. Given some pig images that the model never sees, is there a way to translate pig to cat?
In short, is there a way to translate unseen class to seen class?