GGmorello/FLAMES_only_predicates

该数据集包含多个特征，包括原始索引（original_idx）、输入ID（input_ids）、注意力掩码（attention_mask）和标签（labels）。数据集被分为训练集、验证集和测试集，分别包含2,681,865、335,927和334,533个样本。数据集的下载大小为2,715,895,758字节，总大小为94,684,747,327字节。数据文件路径分别为data/tokenized_split/train-*、data/tokenized_split/val-*和data/tokenized_split/test-*。

The dataset includes multiple features such as original_idx, input_ids, attention_mask, and labels. It is divided into training, validation, and test sets, containing 2,681,865, 335,927, and 334,533 samples respectively. The download size of the dataset is 2,715,895,758 bytes, and the total size is 94,684,747,327 bytes. The data file paths are data/tokenized_split/train-*, data/tokenized_split/val-*, and data/tokenized_split/test-*.