Reubencf/multilingual-image-annotations-text

多语言图像注释（仅文本）是Reubencf/multilingual-image-annotations数据集的纯文本版本，移除了图像和boxed_image列，仅保留文本数据。数据集包含464行数据，支持英语、西班牙语、法语、印地语、中文、阿拉伯语和葡萄牙语。每行数据包含图像ID、英文描述、多语言描述、视觉问答对、对象检测适用性标志、检测结果和模型ID。数据集适用于视觉问答、文本生成、图像字幕生成和对象检测等任务。

Multilingual Image Annotations (Text Only) is the text-only companion to Reubencf/multilingual-image-annotations dataset, with the `image` and `boxed_image` columns removed. It contains 464 rows with annotations in multiple languages (en, es, fr, hi, zh, ar, pt). Each row includes image ID, English description, multilingual descriptions, visual QA pairs, object detection applicability flag, detection results, and model ID. The dataset is suitable for visual question answering, text generation, image captioning, and object detection tasks.