That's the dataset used in the paper.