Closed ChenMnZ closed 1 year ago
Hi, thanks for your interest!
A couple of things:
Euclidian distance is a distance measure, not a similarity measure. So, you need to use -torch.cdist(a,b)
instead.
Of the top of my head, I can't think of anything else.
Thanks! I have successfully reproduced.
Hello, Thank you for your interesting work!
I want to know the implementation of eucl distance function in Tab. 1b. I try to use
scores = torch.cdist(a,b)
to calculate the eucl distance, but the top-1 accuracy on ViT-L/16 with r=8 is only 61.50 by my implementation.Thanks!