3 ms·
The human face training data was probably much more uniform: all well-lit photos of human faces in the center of the photo looking directly at the camera. Where
by rcthompson 8y ago
The human face training data was probably much more uniform: all well-lit photos of human faces in the center of the photo looking directly at the camera. Whereas the training set for this was probably any photos of cats, in any lighting conditions, with the cat in any pose.