I'm trying to find out a dataset with following requirements
(1) It only consists 5~10 subjects (so each subjects has lots of samples to be trained)
(2) More than two kinds of different labels.
(3) The data itself are real valued
e.g.) Facial image datasets of 10 subjects. Each image labeled with ages and race, gender.(So it has three kinds of labels)
Of course it doesn't have to be face images. (for example, hand images of 5~10 subjects, each labeled with race, age gender etc.)
So far, the dataset I found contains large number of subjects but each subject sample size is small, or it has only one kinds of label.
Any help is appreciated