Answering and reasoning questions from text in images. - View it on GitHub