YC-DREAL/ImageNet_Concept_Preference_12k

该数据集包含四个主要字段：被拒绝的信息(rejected)、图片(images)、对话(conversations)和选择的信息(chosen)。被拒绝的信息和选择的信息包括来源(from)和值(value)两个字段，对话字段包含一系列的来源和值，而图片字段则包含图片序列。数据集被划分为训练集(train)，共有11999个示例。数据集主要用于某种涉及图像和对话处理的任务。

The dataset includes four main fields: rejected information, images, conversations, and chosen information. The rejected and chosen information fields consist of from and value subfields, the conversation field contains a series of from and value entries, and the image field includes a sequence of images. The dataset is split into a training set (train) with 11,999 examples. It is intended for tasks involving image and conversation processing.