DT4LM/rte_adversarial_training

该数据集包含三个主要特征：premise（前提）、hypothesis（假设）和label（标签）。premise和hypothesis是字符串类型，表示文本数据，label是整数类型，用于分类或标注。数据集仅包含一个训练集，共有221个样本，文件大小为73742.89652056033字节，下载大小为57148字节。数据集的默认配置文件指定了数据文件的路径为data/train-*。

The dataset contains three main features: premise, hypothesis, and label. Premise and hypothesis are of string type, representing textual data, while label is of integer type, used for classification or annotation. The dataset includes only a training set with 221 examples, a file size of 73742.89652056033 bytes, and a download size of 57148 bytes. The default configuration file for the dataset specifies the data file path as data/train-*.