amuvarma/stt-370k-raw-1dups-re

该数据集包含转录文本和六个facodec序列特征。转录文本以字符串形式存储，而facodec序列特征以int64类型的序列形式存储。数据集仅包含一个训练集，共有370,000个样本，总大小为16,045,716,452字节。数据文件路径为data/train-*。

The dataset contains transcript text and six facodec sequence features. The transcript text is stored as strings, while the facodec sequence features are stored as sequences of int64. The dataset includes only a training set with 370,000 samples, totaling 16,045,716,452 bytes. The data file path is data/train-*.