forked from torch/nn
-
Notifications
You must be signed in to change notification settings - Fork 0
/
SpatialDropout.lua
47 lines (42 loc) · 1.42 KB
/
SpatialDropout.lua
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
local SpatialDropout, Parent = torch.class('nn.SpatialDropout', 'nn.Module')
function SpatialDropout:__init(p)
Parent.__init(self)
self.p = p or 0.5
self.train = true
self.noise = torch.Tensor()
end
function SpatialDropout:updateOutput(input)
self.output:resizeAs(input):copy(input)
if self.train then
if input:dim() == 4 then
self.noise:resize(input:size(1), input:size(2), 1, 1)
elseif input:dim() == 3 then
self.noise:resize(input:size(1), 1, 1)
else
error('Input must be 4D (nbatch, nfeat, h, w) or 3D (nfeat, h, w)')
end
self.noise:bernoulli(1-self.p)
-- We expand the random dropouts to the entire feature map because the
-- features are likely correlated accross the map and so the dropout
-- should also be correlated.
self.output:cmul(torch.expandAs(self.noise, input))
else
self.output:mul(1-self.p)
end
return self.output
end
function SpatialDropout:updateGradInput(input, gradOutput)
if self.train then
self.gradInput:resizeAs(gradOutput):copy(gradOutput)
self.gradInput:cmul(torch.expandAs(self.noise, input)) -- simply mask the gradients with the noise vector
else
error('backprop only defined while training')
end
return self.gradInput
end
function SpatialDropout:setp(p)
self.p = p
end
function SpatialDropout:__tostring__()
return string.format('%s(%f)', torch.type(self), self.p)
end