VictorJuiz/BART_tokenized_cnn_dailymail_3.0.0

该数据集包含三个主要特征：input_ids（整数序列）、attention_mask（8位整数序列）和labels（64位整数序列）。数据集分为训练集、验证集和测试集，分别包含287113、13368和11490个示例。总下载大小为603215391字节，数据集总大小为1528890784字节。数据文件按照分割存储在指定路径下。

The dataset includes three main features: input_ids (sequence of int32), attention_mask (sequence of int8), and labels (sequence of int64). It is divided into training, validation, and test sets, containing 287113, 13368, and 11490 examples respectively. The total download size is 603215391 bytes, and the total dataset size is 1528890784 bytes. Data files are stored in specified paths according to their splits.