mlplo/data/processed/xsum_medium/dataset_dict.json

{"splits": ["train", "validation", "test"]}