fixie-ai/peoples_speech

该数据集包含两个配置：clean和dirty_sa。每个配置都包含以下特征：id（字符串类型）、audio（音频类型，采样率为16000）、duration_ms（整型）、text（字符串类型）和continuation（字符串类型）。数据集被划分为训练集、验证集和测试集，每个划分都有相应的字节数和示例数。clean配置的训练集包含1,501,271个示例，验证集包含18,622个示例，测试集包含34,898个示例。dirty_sa配置的训练集包含548,014个示例，验证集包含18,622个示例，测试集包含34,898个示例。

The dataset includes two configurations: clean and dirty_sa. Each configuration has features such as id, audio, duration_ms, text, and continuation. The dataset is divided into train, validation, and test splits, each with specified number of examples and bytes. The download size and total dataset size are also provided.