0 / 60 seg.

In a sense, we converted one image into billions of training data points, massively reducing the amount of data needed for training.