mlfoundations-dev/hero-1

该数据集包含对话和问题两个主要部分，对话部分包含发送者和信息内容，问题部分是字符串类型的数据。数据集仅有一个训练集分区，包含约846996个示例。数据集的总大小为约12.23GB。

The dataset includes two main sections: conversations and questions. The conversation section contains the sender and the message content, while the question section consists of string-type data. The dataset has only one training set split, containing approximately 846,996 examples. The total size of the dataset is about 12.23GB.