adaadig/TEMP_sweepmetrics

该数据集包含多个音频相关的特征，如文件名、句子ID、字幕文本、超参数值、单词错误率、迭代次数、信噪比、c50值、语音时长、VAD检测到的非语音段、信干比、感知评价质量、短时客观 intelligibility、平均音高、音高标准差、wav2vec自动语音识别字幕、wav2vec的单词错误率、插入、替换和删除错误以及说话人相似度。数据集分为训练集，共有15000个示例。

The dataset includes various audio-related features such as filename, sentence ID, transcript text, hyperparameter values, word error rate, number of iterations, signal-to-noise ratio, c50 value, speech duration, non-speech segments detected by VAD, signal-to-distortion ratio, perceived evaluation quality, short-time objective intelligibility, mean utterance pitch, standard deviation of pitch, wav2vec automatic speech recognition transcript, wav2vec word error rate, insertion, substitution, and deletion errors, as well as speaker similarity. The dataset is split into a training set with a total of 15,000 examples.