amuvarma/ultrachat-25k-audio-flattened-split

该数据集包含多个字段，涉及人类和助手的对话文本以及助手的语音数据。数据集被分割为多个训练集，每个训练集包含318或327个样本，总下载大小约为23MB，数据集总大小约为23.8MB。数据集的字段包括id、human1、assistant1、human2、assistant2、human3、assistant3、assistant_speech1、assistant_speech2和assistant_speech3，其中assistant_speech字段为音频类型。

The dataset contains multiple fields involving dialogue text between humans and assistants, as well as speech data from assistants. The dataset is divided into multiple training sets, each containing 318 or 327 samples, with a total download size of approximately 23MB and a total dataset size of approximately 23.8MB. The fields in the dataset include id, human1, assistant1, human2, assistant2, human3, assistant3, assistant_speech1, assistant_speech2, and assistant_speech3, where the assistant_speech fields are of audio type.