FineGRAIN: Evaluating Failure Modes of Text-to-Image Models with Vision Language Model Judges

Open in new window