Dynamic Multimodal Evaluation with Flexible Complexity by Vision-Language Bootstrapping

Publication
ICLR 2025