A Model for Natural Language Attack on Text Classification and Inference - View it on GitHub