Avalinguo

该数据集名为Avalinguo，包含了1424段非母语英语者的音频记录，这些记录被标注为低、中或高流畅度。在排除了36个语音极少的文件后，剩余1388段录音，采样率为16千赫兹，总时长约2小时，涵盖了对话式口语。这些语音区域根据说话单词的数量，时长介于0.5秒至4.9秒之间。此外，一个包含转写文本和统计分析的标准化版本可在网上获取。该数据集规模为1388段录音，其任务是进行流畅度评估。

This dataset, named Avalinguo, contains 1,424 audio recordings from non-native English speakers, which are annotated with low, medium, or high fluency levels. After excluding 36 files with extremely sparse speech content, 1,388 valid recordings are retained. The recordings have a sampling rate of 16 kHz, with a total duration of roughly 2 hours, and cover conversational spoken English. Each speech segment in the recordings, based on the number of spoken words, has a duration ranging from 0.5 seconds to 4.9 seconds. Additionally, a standardized version including transcriptions and statistical analyses is publicly available online. The dataset comprises 1,388 recordings, with the associated task being fluency assessment.