mlfoundations-dev/oh_v3.1_wo_evolinstruct

该数据集主要包含对话数据，每个对话记录包括来源（from）和内容（value）两个主要部分。此外，数据集还包含一个名为source_label_exact的字段，用于存储字符串序列。数据集分为训练集，包含947,056个样本，总数据大小为1,835,194,639字节，下载大小为966,301,017字节。

This dataset primarily contains dialogue data, with each dialogue record including two main parts: the source (from) and the content (value). Additionally, the dataset includes a field named source_label_exact for storing sequences of strings. The dataset is divided into a training set, containing 947,056 samples, with a total data size of 1,835,194,639 bytes and a download size of 966,301,017 bytes.