Jiwon-Kang/OCR-Synthetic-Rendered-200K

这是一个多模态视觉问答（VQA）数据集，包含图像、问题和答案的配对数据。数据集特征包括图像、问题文本、答案文本、来源数据集、来源分割、来源索引、渲染文本、模板、宽度和高度。它适用于训练和评估视觉问答模型，支持从图像中理解内容并回答相关问题。数据集分为训练集（200,000个样本）和验证集（500个样本），总大小约7.95 GB。

This is a multimodal visual question answering (VQA) dataset containing paired data of images, questions, and answers. The dataset features include image, question text, answer text, source dataset, source split, source index, render text, template, width, and height. It is suitable for training and evaluating visual question answering models, supporting content understanding from images and answering related questions. The dataset is divided into a training set (200,000 examples) and a validation set (500 examples), with a total size of approximately 7.95 GB.