Skip to content

partial con #39

Description

@SWPU-WuDan

some questions about "class PartialConv2d(nn.Conv2d): "
firstly, what's the purpose of " self.mask_ratio = torch.mul(self.mask_ratio, self.update_mask)"? mabe the is the update_mask which is use for the next layer,so,why, torch.mul(self.mask_ratio, self.update_mask)
secendly, what's the purpose of "output = torch.mul(output, self.update_mask)" the next feature layer is the feature with mask?
lastly, there is no computition about filter weight matrix and X (WX).so when type "class PartialConv2d(nn.Conv2d)",the code will first jump to nn.Conv2d ?

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions