Enhancing Robustness of Vision-Language Models through Orthogonality Learning and Cross-Regularization

VLMRe-assign community