ASCIIEval/ASCIIEval

这是一个英文数据集，适用于视觉问答、文本生成、图像文本到文本和问答等任务。数据集规模较小，小于5K。它具有多模态特性，包含字符画，并支持大型语言模型和多语言大型语言模型。此外，该数据集在视觉感知方面有所涉及，并被用作一个基准数据集。

This is an English dataset suitable for tasks such as visual question answering, text generation, image-text-to-text, and question answering. The dataset is small, with fewer than 5K entries. It is multimodal, includes ascii art, and supports large language models and multilingual large language models. Additionally, the dataset is related to visual perception and is used as a benchmark.