Turbo-AI/train-multi-negatives

该数据集包含三个主要特征：id（数据类型为int64）、text（数据类型为string）和relevant（包含id和text两个子特征，数据类型分别为int64和string）。此外，还有一个not_relevant特征，其结构与relevant相同。数据集包含一个名为train的分割，该分割包含118,952个示例，总大小为1,185,738,302字节。下载大小为392,995,499字节，数据集总大小为1,185,738,302字节。配置信息显示，数据文件位于data/train-*路径下。

The dataset contains three main features: id (data type int64), text (data type string), and relevant (which includes two sub-features: id and text, with data types int64 and string respectively). Additionally, there is a not_relevant feature with the same structure as relevant. The dataset includes a split named train, which contains 118,952 examples with a total size of 1,185,738,302 bytes. The download size is 392,995,499 bytes, and the total dataset size is 1,185,738,302 bytes. The configuration information indicates that the data files are located under the path data/train-*.