minato-ryan/emilia-cjk-xcodec2

该数据集包含多个配置（ja_long，ja_short，ja_skipped，ja_very_long），每个配置都有如uid、dnsmos、时长、语言、说话人、文本、wav文件和ids等特征。每个配置的训练分割都包含了示例数量和字节数的信息。此外，还提供了每个配置的下载大小和数据集大小。

The dataset comprises multiple configurations (ja_long, ja_short, ja_skipped, ja_very_long), each with features like uid, dnsmos, duration, language, speaker, text, wav file, and ids. Each configurations train split includes information on the number of examples and bytes. Additionally, the download size and dataset size for each configuration are provided.