burkimbia/speech-dataset-rejected

该数据集是一个包含音频和文本数据的集合，主要用于语音处理或自然语言处理任务。数据集包含4163个训练样本，每个样本具有文本内容、音频文件、持续时间、说话者信息、数据来源、性别标识、保留标志、重复标志、原因列表、剪辑标志、静音比例、字符每秒速率、单词计数和检测语言等特征字段。这些特征支持音频质量分析、说话者识别、语言检测和多模态应用。

This dataset is a collection of audio and text data, primarily designed for speech processing or natural language processing tasks. It includes 4163 training examples, each featuring text content, audio files, duration, speaker information, data source, gender identifier, keep flag, duplicate flag, reasons list, clipped flag, silence ratio, characters per second rate, word count, and detected language. These features facilitate audio quality analysis, speaker identification, language detection, and multimodal applications.