Towards Efficient Adversarial Training on Vision Transformers | Read Paper on Bytez