Sample-agnostic Adversarial Perturbation for Vision-Language Pre-training Models | Read Paper on Bytez