lion-ai/bigos

该数据集包含音频和转录文本两个主要特征，音频数据类型为音频，转录文本数据类型为字符串。数据集分为训练集和验证集，训练集包含82025个样本，总大小为38444026358字节；验证集包含14254个样本，总大小为5360001471字节。数据集总大小为43804027829字节，下载大小为29516638923字节。配置文件指定了训练集和验证集数据文件的路径。

The dataset includes two main features: audio and transcription, with audio data type and string data type respectively. It is divided into train and validation splits. The train split contains 82,025 samples with a total size of 38,444,026,358 bytes, and the validation split contains 14,254 samples with a total size of 5,360,001,471 bytes. The total dataset size is 43,804,027,829 bytes, and the download size is 29,516,638,923 bytes. The configuration specifies the paths for the data files corresponding to each split.