jan-hq/gigaspeech-vie-sub

该数据集包含音频ID、转录文本和音频数据三个特征。数据集仅包含一个训练集，共有52,081个样本，总大小为7,723,539,252.875字节，下载大小为7,705,893,691字节。配置文件中指定了默认配置，数据文件路径为data/train-*。

The dataset contains three features: audio_id (audio ID, data type is string), transcription (transcription text, data type is string), and audio (audio data, data type is audio). The dataset includes only a training set with 52,081 samples, totaling 7,723,539,252.875 bytes in size, and a download size of 7,705,893,691 bytes. The configuration file specifies the default configuration, with the data file path being data/train-*.