ZiqinZhou66

Results 1 issues of ZiqinZhou66

When I changed the pretrained model fro ViT-B/32 to ViT-B/16 and ViT-L/14, I could not get such an attention map. Do you have the same results?