mlfoundations-dev/top_7_ranking_stackexchange

该数据集包含了指令、完成情况、原始行索引以及对话信息。指令和完成情况以字符串形式存储，原始行索引为整型，对话信息包括对话来源和内容。数据集划分为训练集，共有350000个示例，总大小约为2.4GB。提供的默认配置指定了训练数据的文件路径。

The dataset includes instructions, completions, original row indices, and conversation information. Instructions and completions are stored as strings, the original row index is an integer, and conversation information includes the source and content of the conversation. The dataset is split into a training set with a total of 350,000 examples, totaling approximately 2.4GB. The default configuration specifies the file path for the training data.