mteb/svq

Simple Voice Questions (SVQ) 是一个包含17种语言、26个地区的短音频问题的数据集，这些音频在多种音频条件下录制。数据收集过程中，说话者在四种不同环境下（安静环境、背景语音噪声、交通噪声、媒体噪声）使用自己的设备录制音频。查询文本来源于XTREME-UP的检索和问答基准数据集的验证和测试集。数据集包含700个独特的说话者，记录了说话者的性别（女性、男性、非二元、无回答）和年龄信息。数据集未预先分割为训练、验证或测试子集，而是作为一个完整的集合发布，用户需要自行设计分割策略。

Simple Voice Questions (SVQ) is a set of short audio questions recorded in 26 locales across 17 languages under multiple audio conditions. Speakers recorded the audio in four different environments (clean, background speech noise, traffic noise, media noise) using their own devices. The querys text comes from validation and test sets of the XTREME-UPs retrieval and question answering benchmark datasets. The dataset includes 700 unique speakers, with recorded speaker gender (female, male, non_binary, no_answer) and age information. The dataset is not pre-divided into training, validation, or testing subsets but is released as a complete collection, requiring users to design their own splitting strategies.