trl-internal-testing/zen-multi-image

该数据集包含三种配置：对话语言模型、对话提示完成和对话仅提示。每种配置都包括消息特征，其中包含文本和类型字段以及角色信息。某些配置还包括图像。数据集被分为训练集和测试集，并提到了各自的示例数量和文件大小。每个配置的下载大小和总数据集大小也已提供。

The dataset consists of three configurations: conversational language modeling, conversational prompt completion, and conversational prompt only. Each configuration includes features for messages, which contain text and type fields, along with role information. Some configurations also include images. The dataset is split into training and testing sets, with the respective number of examples and file sizes mentioned. The download size and total dataset size are provided for each configuration.