Are the outputs of Dropout2d scaled?

vision
1

The outputs of Dropout are scaled by 1 / (1 - p) to make the layer indentity during testing. Is it same case for Dropout2d?

2

Yes, seems to be the case as seen here:

drop = nn.Dropout2d()
x = torch.ones(2, 2, 2, 2)
out = drop(x)
print(out)
> tensor([[[[0., 0.],
          [0., 0.]],

         [[0., 0.],
          [0., 0.]]],


        [[[2., 2.],
          [2., 2.]],

         [[0., 0.],
          [0., 0.]]]])
1 Like
3

Thanks. So stupied of me not running some code .