Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Can the SparseCenterLoss be implemented in another way? #11

Open
cotyyang opened this issue Dec 21, 2022 · 0 comments
Open

Can the SparseCenterLoss be implemented in another way? #11

cotyyang opened this issue Dec 21, 2022 · 0 comments

Comments

@cotyyang
Copy link

cotyyang commented Dec 21, 2022

hello,your code loss.py and center_loss_pytorch are similar,but center loss has another implement,KaiyangZhou/pytorch-center-loss#20 ,therefore, can the scheme be modified as follows?

class CenterLoss(nn.Module):
    def __init__(self, num_class=10, num_feature=2):
        super(CenterLoss, self).__init__()
        self.num_class = num_class
        self.num_feature = num_feature
        self.centers = nn.Parameter(torch.randn(self.num_class, self.num_feature))

    def forward(self, feat, A, label):
        center = self.centers[label]
        dist = (A * ((feat - center).pow(2))).sum(dim=-1)
        loss = torch.clamp(dist, min=1e-12, max=1e+12).mean(dim=-1)

        return loss
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant