ylacombe/peoples_speech-tags-annotated

该数据集包含两个配置（clean和default），主要涉及语音处理和文本分析。特征字段包括语音时长、音高均值、信噪比、语音清晰度、语音质量评估指标（如STOI、SI-SDR、PESQ等）、噪声类型、混响情况、语音单调性、原始文本、处理后的文本以及性别信息。数据集分为训练集、验证集和测试集，分别包含1501271、18622和34898个样本。

This dataset includes two configurations (clean and default) and is primarily related to speech processing and text analysis. The feature fields include speech duration, mean pitch, signal-to-noise ratio, speech intelligibility, speech quality evaluation metrics (such as STOI, SI-SDR, PESQ), noise type, reverberation, speech monotony, original text, processed text, and gender information. The dataset is divided into training, validation, and test sets, containing 1501271, 18622, and 34898 samples respectively.