1 paper
Rob Reijtenbach, Suzan Verberne, Gijs Wijnholds
In this paper we present and validate a new synthetic dataset for training visual entailment models. Existing datasets for visual entailment are small and sparse compared to datase…