Hi, Davis, have you trained the face recognition with ms-celeb-1m dataset? they have 3 types of data: the original one, the cropped and aligned one by their tool, my question is: which one is better if I want to train the resnet for face recognition within dlib? thank you so much.
Hi, Davis, have you trained the face recognition with ms-celeb-1m dataset? they have 3 types of data: the original one, the cropped and aligned one by their tool, my question is: which one is better if I want to train the resnet for face recognition within dlib? thank you so much.