amuvarma/100k-1cols-stt-textloss

该数据集包含100,000个训练样本，每个样本包含input_ids、attention_mask和labels三个特征。input_ids是int32类型的序列，attention_mask是int8类型的序列，labels是int64类型的序列。数据集总大小为2,663,600,000字节，下载大小为219,494,143字节。数据集的默认配置指定了训练数据文件的路径。

The dataset contains 100,000 training samples, each with three features: input_ids, attention_mask, and labels. input_ids is a sequence of int32, attention_mask is a sequence of int8, and labels is a sequence of int64. The total size of the dataset is 2,663,600,000 bytes, and the download size is 219,494,143 bytes. The default configuration of the dataset specifies the path to the training data files.