AdoCleanCode/commonvoice_en_mfa_correct

该数据集是一个包含音频文件及其转录文本、单词和音素时间戳信息的语音数据集。具体特征包括：id（字符串）、speaker_id（字符串）、音频（采样率16kHz）、转录文本（字符串）、单词列表（每个单词包含单词文本、开始时间和结束时间）和音素列表（每个音素包含音素文本、开始时间和结束时间）。数据集仅包含训练集，共有1,071,533个样本，总大小为176,560,097,408.375字节。

This dataset is a speech dataset containing audio files along with their transcriptions, word-level and phoneme-level timestamps. The features include: id (string), speaker_id (string), audio (sampling rate 16kHz), transcript (string), words list (each word includes the word text, start time, and end time), and phonemes list (each phoneme includes the phoneme text, start time, and end time). The dataset only contains a training split with 1,071,533 examples and a total size of 176,560,097,408.375 bytes.