LuigiRiz / NOPS

Official implementation of the CVPR 2023 paper "Novel Class Discovery for 3D Point Cloud Semantic Segmentation"
59 stars 5 forks source link

Sinkhorn Algorithm #5

Open kleinzcy opened 1 year ago

kleinzcy commented 1 year ago

Thanks your code.

It seems that there is something wrong in your Sinkhorn Algorithm.

        features = torch.nn.functional.normalize(features, dim=1, p=2)
        head = torch.nn.functional.normalize(head, dim=1, p=2)

In my understanding, the dim of "head" is $d\times N$. It is not right to normalize along the dim 1. The right way is to normalize the head along with dim 0. Do I understand right?