VCR

The dataset used in the paper for visual commonsense reasoning.

BibTex: