mlfoundations-dev/oh_v3.1_wo_unnatural_instructions

该数据集包含对话数据，每个对话包含from和value两个字段，分别表示对话的来源和内容。此外，数据集还包含一个名为source_label_exact的序列字段。数据集分为训练集，包含1,001,107个样本，总大小为1,995,863,042字节。数据集的下载大小为1,049,277,322字节。

The dataset contains conversation data, with each conversation including from and value fields, representing the source and content of the conversation, respectively. Additionally, the dataset includes a sequence field named source_label_exact. The dataset is divided into a training set, containing 1,001,107 samples, with a total size of 1,995,863,042 bytes. The download size of the dataset is 1,049,277,322 bytes.