PicT: A Slim Weakly Supervised Vision Transformer for Pavement Distress Classification | Read Paper on Bytez