It is a deep learning framework intended for image classification and segmentation. It has features like simple CPU and GPU switching, optimised model definition and configuration, computation ...