This is a model used to predict the presence of disease in a plant and also classifies the disease. Several CNN models like Resnet, InceptionV3, Densenet were used in this case and the final model which is an ensemble of three models of the models was used to yield a ROC-AUC of 0.97.

Dataset - https://www.kaggle.com/c/plant-pathology-2020-fgvc7/data

Private Score(50% of test data) is shown in the middle while Public Score(remaining 50% of test data) is shown on far right.