13 ms·
Attention is more nuanced, operating at different scales. A common flaw in deep fake faces is pupils that aren't quite round, so this type of model could be use
by robbedpeter 5y ago
Attention is more nuanced, operating at different scales. A common flaw in deep fake faces is pupils that aren't quite round, so this type of model could be used to target errors like that. Or if you had a picture of a library bookshelf, the model can attend to the details of individual books, and also attend to the letters of their titles at a per book level.