OUTCOMESAI/medical_speech_corpus

该数据集包含两个配置：en和zh-en。每个配置都有相同的特征，包括term、text、audio、english_text、voice和type。audio特征的采样率为16000。数据集分为训练集和测试集，分别提供了字节数和样本数。en配置的数据集较大，包含366732个训练样本和107863个测试样本。zh-en配置的数据集较小，包含19944个训练样本和4986个测试样本。

The dataset includes two configurations: en and zh-en. Each configuration has the same features, including term, text, audio, english_text, voice, and type. The audio feature has a sampling rate of 16000. The dataset is divided into training and test sets, with the number of bytes and examples provided for each. The en configuration has a larger dataset, containing 366732 training samples and 107863 test samples. The zh-en configuration has a smaller dataset, containing 19944 training samples and 4986 test samples.