> 820 is an incredibly small data size for deep learning Whether or not 820 samples is little or a lot depends on the nature of the problem and the amount of i

> 820 is an incredibly small data size for deep learning Whether or not 820 samples is little or a lot depends on the nature of the problem and the amount of information contained in each sample. For example, in 3D medical imagining we routinely reach superhuman performance with training sets of a couple of a hundred images (training from random weights, i.e. with no pretraining).