COCO Images: Training images [83K/13GB], Validation Images [41K/6GB]
VQA-E dataset: Google Drive
Annotation format:
annotation{
"image_id" : int,
"question_type" : str,
"answer_type" : str,
"answers" : [answer],
"multiple_choice_answer" : str,
"explanation": [explanation, score]
}
If you use the VQA-E dataset as part of any published research, please acknowledge the following paper
@@article{li2018vqae,
title={VQA-E: Explaining, Elaborating, and Enhancing Your Answers for Visual Questions},
author={Li, Qing and Tao, Qingyi and Joty, Shafiq and Cai, Jianfei and Luo, Jiebo},
journal={ECCV},
year={2018}
}
MIT License.