VQAvs

VQAvs is a dataset for visual question answering, containing questions answerable using visual images.

BibTex: