Mixup Data Augmentation - Deep Learning Dictionary
Mixup Data Augmentation Technique - Deep Learning Dictionary
Mixup is a specific type of data augmentation technique that differs from many of the basic data augmentation implementations.
Suppose we have two classes of images for which we'll train a network. Mixup creates new samples to include in the training set by combining (i.e. mixing up) the existing samples and the corresponding labels.
Committed by on