CambioMoney/ami-speaker-analysis_full_run_deepgram_5

该数据集包含会议音频及其相关文本信息，每个样本包括会议ID、音频ID、文本内容、音频数据（包含音频数组、路径和采样率）、开始时间、结束时间、麦克风ID、说话者ID、是否完整以及是否为原始片段等信息。数据集分为训练集、验证集和测试集，分别包含163228、18847和18227个样本。数据集的下载大小为13503746937字节，总大小为59003764821字节。

This is a dataset containing meeting records, featuring multiple attributes such as meeting ID, audio ID, text, audio information (including audio array, path, and sampling rate), start time, end time, microphone ID, speaker ID, completeness, and original segment identification. The dataset is divided into training, validation, and test sets, each containing different numbers of samples and bytes.