imb-cls-mask-distilled-libri-one-sec-cv12/chunk_182

该数据集包含音频和浮点数序列两种特征。音频特征的采样率为16000Hz，浮点数序列特征用于存储模型的输出logits。数据集包含一个训练集分割，共有5660个样本，总大小为181454647.56308517字节。数据集的下载大小为161756683字节。默认配置下的数据文件路径为data/train-*。

This dataset contains two main features: audio and logits. The audio feature is of audio type with a sampling rate of 16000Hz, and the logits feature is a sequence of float32 values. The dataset includes a training split with 5660 examples, totaling 181454647.56308517 bytes in size. The download size of the dataset is 161756683 bytes. The default configuration specifies the data file path as data/train-*.