Improving Generative Visual Dialog by Answering Diverse Questions

Open in new window