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
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
|
local Module = torch.class('nn.Module')
function Module:__init()
self.gradInput = torch.Tensor()
self.output = torch.Tensor()
end
function Module:parameters()
if self.weight and self.bias then
return {self.weight, self.bias}, {self.gradWeight, self.gradBias}
elseif self.weight then
return {self.weight}, {self.gradWeight}
elseif self.bias then
return {self.bias}, {self.gradBias}
else
return
end
end
function Module:updateOutput(input)
return self.output
end
function Module:forward(input)
return self:updateOutput(input)
end
function Module:backward(input, gradOutput, scale)
scale = scale or 1
self:updateGradInput(input, gradOutput)
self:accGradParameters(input, gradOutput, scale)
return self.gradInput
end
function Module:backwardUpdate(input, gradOutput, lr)
self:updateGradInput(input, gradOutput)
self:accUpdateGradParameters(input, gradOutput, lr)
return self.gradInput
end
function Module:updateGradInput(input, gradOutput)
return self.gradInput
end
function Module:accGradParameters(input, gradOutput, scale)
end
function Module:accUpdateGradParameters(input, gradOutput, lr)
local gradWeight = self.gradWeight
local gradBias = self.gradBias
self.gradWeight = self.weight
self.gradBias = self.bias
self:accGradParameters(input, gradOutput, -lr)
self.gradWeight = gradWeight
self.gradBias = gradBias
end
function Module:sharedAccUpdateGradParameters(input, gradOutput, lr)
if self:parameters() then
self:zeroGradParameters()
self:accGradParameters(input, gradOutput, 1)
self:updateParameters(lr)
end
end
function Module:zeroGradParameters()
local _,gradParams = self:parameters()
if gradParams then
for i=1,#gradParams do
gradParams[i]:zero()
end
end
end
function Module:updateParameters(learningRate)
local params, gradParams = self:parameters()
if params then
for i=1,#params do
params[i]:add(-learningRate, gradParams[i])
end
end
end
function Module:training()
self.train = true
end
function Module:evaluate()
self.train = false
end
function Module:share(mlp, ...)
local arg = {...}
for i,v in ipairs(arg) do
if self[v] ~= nil then
self[v]:set(mlp[v])
self.accUpdateGradParameters = self.sharedAccUpdateGradParameters
mlp.accUpdateGradParameters = mlp.sharedAccUpdateGradParameters
end
end
return self
end
function Module:clone(...)
local f = torch.MemoryFile("rw"):binary()
f:writeObject(self)
f:seek(1)
local clone = f:readObject()
f:close()
if select('#',...) > 0 then
clone:share(self,...)
end
return clone
end
function Module:type(type)
-- find all tensors and convert them
for key,param in pairs(self) do
if torch.typename(param) and torch.typename(param):find('torch%..+Tensor') then
self[key] = param:type(type)
end
end
-- find submodules in classic containers 'modules'
if self.modules then
for _,module in ipairs(self.modules) do
module:type(type)
end
end
return self
end
function Module:float()
return self:type('torch.FloatTensor')
end
function Module:double()
return self:type('torch.DoubleTensor')
end
function Module:cuda()
return self:type('torch.CudaTensor')
end
function Module:reset()
end
function Module:getParameters()
-- get parameters
local parameters,gradParameters = self:parameters()
local function storageInSet(set, storage)
local storageAndOffset = set[torch.pointer(storage)]
if storageAndOffset == nil then
return nil
end
local storage, offset = unpack(storageAndOffset)
return offset
end
-- this function flattens arbitrary lists of parameters,
-- even complex shared ones
local function flatten(parameters)
local Tensor = parameters[1].new
local storages = {}
local nParameters = 0
for k = 1,#parameters do
local storage = parameters[k]:storage()
if not storageInSet(storages, storage) then
storages[torch.pointer(storage)] = {storage, nParameters}
nParameters = nParameters + storage:size()
end
end
local flatParameters = Tensor(nParameters):fill(1)
local flatStorage = flatParameters:storage()
for k = 1,#parameters do
local storageOffset = storageInSet(storages, parameters[k]:storage())
parameters[k]:set(flatStorage,
storageOffset + parameters[k]:storageOffset(),
parameters[k]:size(),
parameters[k]:stride())
parameters[k]:zero()
end
local cumSumOfHoles = flatParameters:float():cumsum(1)
local nUsedParameters = nParameters - cumSumOfHoles[#cumSumOfHoles]
local flatUsedParameters = Tensor(nUsedParameters)
local flatUsedStorage = flatUsedParameters:storage()
for k = 1,#parameters do
local offset = cumSumOfHoles[parameters[k]:storageOffset()]
parameters[k]:set(flatUsedStorage,
parameters[k]:storageOffset() - offset,
parameters[k]:size(),
parameters[k]:stride())
end
for _, storageAndOffset in pairs(storages) do
local k, v = unpack(storageAndOffset)
flatParameters[{{v+1,v+k:size()}}]:copy(Tensor():set(k))
end
if cumSumOfHoles:sum() == 0 then
flatUsedParameters:copy(flatParameters)
else
for k = 1,flatUsedParameters:nElement() do
flatUsedParameters[k] = flatParameters[k+cumSumOfHoles[k]]
end
end
return flatUsedParameters
end
-- flatten parameters and gradients
local flatParameters = flatten(parameters)
local flatGradParameters = flatten(gradParameters)
-- return new flat vector that contains all discrete parameters
return flatParameters, flatGradParameters
end
function Module:__call__(input, gradOutput)
self:forward(input)
if gradOutput then
self:backward(input, gradOutput)
return self.output, self.gradInput
else
return self.output
end
end
|