size_categories: - n<1K
This dataset contains six files at the root, four for the training split, and two for the test split.
