require 'dagnn'
+function checkGrad(model, criterion, input, target)
+ local params, gradParams = model:getParameters()
+
+ local epsilon = 1e-5
+
+ local output = model:forward(input)
+ local loss = criterion:forward(output, target)
+ local gradOutput = criterion:backward(output, target)
+ gradParams:zero()
+ model:backward(input, gradOutput)
+ local analyticalGradParam = gradParams:clone()
+
+ for i = 1, params:size(1) do
+ local x = params[i]
+
+ params[i] = x - epsilon
+ local output0 = model:forward(input)
+ local loss0 = criterion:forward(output0, target)
+
+ params[i] = x + epsilon
+ local output1 = model:forward(input)
+ local loss1 = criterion:forward(output1, target)
+
+ params[i] = x
+
+ local ana = analyticalGradParam[i]
+ local num = (loss1 - loss0) / (2 * epsilon)
+ local err = torch.abs(num - ana) / torch.abs(num)
+
+ print(
+ err .. ' checkGrad ' .. i
+ .. ' analytical ' .. ana
+ .. ' numerical ' .. num
+ )
+ end
+
+end
+
+function printTensorTable(t)
+ if torch.type(t) == 'table' then
+ for i, t in pairs(t) do
+ print('-- ELEMENT [' .. i .. '] --')
+ printTensorTable(t)
+ end
+ else
+ print(tostring(t))
+ end
+end
+
+-- torch.setnumthreads(params.nbThreads)
+torch.setdefaulttensortype('torch.DoubleTensor')
+torch.manualSeed(2)
+
+-- +--> c ----> e --+
+-- / / \
+-- / / \
+-- input --> a --> b ---> d ----+ g --> output
+-- \ /
+-- \ /
+-- +--> f ---+
+
a = nn.Linear(10, 10)
b = nn.ReLU()
c = nn.Linear(10, 3)
d = nn.Linear(10, 3)
e = nn.CMulTable()
-f = nn.Linear(3, 2)
+f = nn.Linear(3, 3)
+g = nn.CAddTable()
---[[
+----------------------------------------------------------------------
- a -----> b ---> c ----> e ---
- \ /
- \--> d ---/
- \
- \---> f ---
-]]--
+model = nn.DAG()
-g = DAG:new()
+model:addEdge(a, b)
+model:addEdge(b, c)
+model:addEdge(b, d)
+model:addEdge(c, e)
+model:addEdge(d, e)
+model:addEdge(d, f)
+model:addEdge(e, g)
+model:addEdge(f, g)
-g:setInput(a)
-g:setOutput({ e, f })
-g:addEdge(c, e)
-g:addEdge(a, b)
-g:addEdge(d, e)
-g:addEdge(b, c)
-g:addEdge(b, d)
-g:addEdge(d, f)
+model:setInput(a)
+model:setOutput(g)
-g:order()
+input = torch.Tensor(3, 10):uniform()
-g:print(graph)
+print('******************************************************************')
+print('** updateOutput **************************************************')
+print('******************************************************************')
-input = torch.Tensor(3, 10):uniform()
+output = model:updateOutput(input):clone()
+
+printTensorTable(output)
+
+print('******************************************************************')
+print('** updateGradInput ***********************************************')
+print('******************************************************************')
+
+gradInput = model:updateGradInput(input, output)
+
+printTensorTable(gradInput)
+
+print('******************************************************************')
+print('** checkGrad *****************************************************')
+print('******************************************************************')
-output = g:updateOutput(input)
+output:uniform()
-print(output[1])
-print(output[2])
+checkGrad(model, nn.MSECriterion(), input, output)