Open-ended VQA benchmarking of Vision-Language models by exploiting Classification datasets and their semantic hierarchy.

Simon Ging, Mara Alejandra Bravo, Thomas Brox

Browse the full ICLR paper archive.