mm-eval/VisIT-Bench
收藏官方服务:
资源简介:
该数据集是一个包含图像和文本消息的多模态数据集,主要用于测试目的。数据集包括578个示例,每个示例包含一个唯一的ID、一个图像列表(media)和一个字符串消息(messages)。数据以测试分割形式提供,总大小约为161MB,下载大小相近。数据集未提供具体应用场景或来源描述,但结构表明可能用于图像-文本交互任务,如对话系统或视觉问答。
This dataset is a multimodal dataset containing images and text messages, primarily intended for testing purposes. It includes 578 examples, each consisting of a unique ID, a list of images (media), and a string message (messages). The data is provided in a test split format, with a total size of approximately 161MB and a similar download size. The dataset does not specify application scenarios or sources, but its structure suggests potential use in image-text interaction tasks, such as dialogue systems or visual question answering.
提供机构:
mm-eval


