threll-ai/common_voice_22_0

该数据集包含两个不同配置（nn-NO和nb-NO）的音频片段和相关元数据。每个音频片段都有采样率为48000的音频文件、文本内容（sentence）、投票信息（up_votes和down_votes）、年龄、性别、口音、地区、段落、变体和连续性标记。数据集分为训练集、验证集和测试集，其中nn-NO配置的训练集有464个示例，验证集有405个示例，测试集有423个示例；nb-NO配置的训练集有227个示例，验证集有33个示例，测试集有116个示例。

The dataset consists of two configurations (nn-NO and nb-NO) of audio clips along with associated metadata. Each audio clip includes an audio file with a sampling rate of 48000, text content (sentence), voting information (up_votes and down_votes), age, gender, accent, locale, segment, variant, and continuation tags. The dataset is split into training, validation, and test sets, with the nn-NO configuration having 464 examples in the training set, 405 in the validation set, and 423 in the test set; the nb-NO configuration has 227 examples in the training set, 33 in the validation set, and 116 in the test set.