Supplementary Material for DropPos: Pre-Training Vision Transformers by 1 Reconstructing Dropped Positions