speedcell4/ec40-spm

该数据集包含了两个文本序列（text1和text2），两种语言类型（lang1和lang2），以及两个文本大小（size1和size2）。数据集分为训练集（train）、验证集（dev）、测试集（test）和一个名为zero的集合，每个集合都有各自的数据量和大小。数据集支持默认配置，可通过指定路径加载不同集合的数据文件。

The dataset includes two text sequences (text1 and text2), two language types (lang1 and lang2), and two text sizes (size1 and size2). It is divided into training set (train), validation set (dev), test set (test), and a set named zero, each with its own data volume and size. The dataset supports a default configuration and can load data files from different sets by specifying the path.