WuH1n/WenetSpeech4TTS-Tokenized

该数据集包含音频及其对应文本，以及一些元数据信息如唯一标识符、音频质量评分和事件时间戳。具体来说，每个样本都包括一个字符串类型的唯一标识符（utt_id），一个采样率为16000Hz的音频文件（audio），一段文本（text），一个浮点数表示的音频质量评分（dnsmos），以及一个字符串类型的时间戳（timestamp）。数据集目前只有一个训练部分，名为train.Premium，包含100个样本。整个数据集的大小为28034900字节，下载大小为25932515字节。

The dataset includes audio files with their corresponding text, as well as metadata such as unique identifiers, audio quality scores, and timestamps. Specifically, each sample consists of a string-type unique identifier (utt_id), an audio file with a sampling rate of 16000Hz (audio), a text segment (text), a floating-point number representing the audio quality score (dnsmos), and a string-type timestamp (timestamp). The dataset currently has only one training part named train.Premium, containing 100 samples. The total size of the dataset is 28034900 bytes, with a download size of 25932515 bytes.