This repository contains the dataset used in the paper "RealCQA: Scientific Chart Question Answering as a Test-Bed for First-Order Logic" (ICDAR 2023). The dataset is designed to facilitate research in complex question answering, involving a diverse set of real-world images and associated textual question-answer pairs.
The RealCQA dataset consists of 28,266 images, and corresponding 2 million… See the full description on the dataset page:
https://huggingface.co/datasets/sal4ahm/RealCQA.
