Bean leaf disease (augmented ibean) dataset

This is an augmented dataset of the ibean dataset originally published by Makerere's AI lab. This augmented set expands the training and validation sets 40-times. Like the original dataset, this augmented dataset is nearly balanced across the classes. It contains a total of 41,300 training images; 13,800 angular leaf spot images, 13,900 bean rust images and 13,600 healthy leaves, and a total of 5,320 validation images (1760 angular leaf spot images, 1800 bean rust images and 1760 healthy leaves images. To allow for a fair comparison across previous models, the test set remains the 128images provided in the original dataset.