kanishka/babylm2-subset

该数据集包含文本数据，主要分为训练集和验证集两部分。训练集包含8,434,896个样本，总大小为380,752,170字节；验证集包含834,155个样本，总大小为40,660,051字节。整个数据集的下载大小为235,635,151字节，总大小为421,412,221字节。数据文件按照分割存储在指定的路径下。

This dataset contains text data, primarily divided into a training set and a validation set. The training set includes 8,434,896 samples with a total size of 380,752,170 bytes; the validation set includes 834,155 samples with a total size of 40,660,051 bytes. The entire dataset has a download size of 235,635,151 bytes and a total size of 421,412,221 bytes. Data files are stored in specified paths according to their splits.