opedromartins/speaker-datasets

该数据集包含五个配置不同的子数据集：cml、common-voice、mls、tedx和vt。每个数据集都包含音频文件名、音频数据、发言者信息和语言类型等特征。其中，common-voice和vt数据集中的音频特征还提供了采样率和解码信息。这些数据集都被划分为训练集，并提供了示例数量和文件大小的信息。同时，还列出了每个配置的数据文件路径。

The dataset consists of five sub-datasets with different configurations: cml, common-voice, mls, tedx, and vt. Each dataset includes features such as audio filename, audio data, speaker information, and language type. The audio feature in the common-voice and vt datasets also provides sampling rate and decoding information. These datasets are split into training sets with information on the number of examples and file size. Additionally, the paths to the data files for each configuration are listed.