shchoi1019/256token

该数据集包含多个特征，包括input_ids、token_type_ids、attention_mask和labels，这些特征分别以int32、int8、int8和int64的数据类型表示。数据集主要包含一个训练集分割，该分割包含63,672,428个示例，总大小为229,220,740,800字节。数据集的下载大小为51,872,472,203字节，实际数据集大小为229,220,740,800字节。配置文件中指定了训练数据文件的路径。

The dataset includes several features such as input_ids, token_type_ids, attention_mask, and labels, represented by data types int32, int8, int8, and int64 respectively. It primarily contains a training split with 63,672,428 examples, totaling 229,220,740,800 bytes. The download size of the dataset is 51,872,472,203 bytes, and the actual dataset size is 229,220,740,800 bytes. The configuration file specifies the path to the training data files.