require 'dagnn'
+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)
e = nn.CMulTable()
f = nn.Linear(3, 2)
---[[
-
- a -----> b ---> c ----> e ---
- \ /
- \--> d ---/
- \
- \---> f ---
-]]--
-
-g = nn.DAG:new()
+-- a -----> b ---> c ----> e ---
+-- \ /
+-- \--> d ---/
+-- \
+-- \---> f ---
-g:setInput(a)
-g:setOutput({ e })
+g = nn.DAG()
g:addEdge(c, e)
g:addEdge(a, b)
g:addEdge(d, e)
g:addEdge(b, c)
g:addEdge(b, d)
--- g:addEdge(d, f)
+g:addEdge(d, f)
--- g = torch.load('dag.t7')
+g:setInput({{a}})
+g:setOutput({ e, f })
g:print()
input = torch.Tensor(3, 10):uniform()
-output = g:updateOutput(input)
+output = g:updateOutput({{ input }})
-if torch.type(output) == 'table' then
- for i, t in pairs(output) do
- print(tostring(i) .. ' -> ' .. tostring(t))
- end
-else
- print(tostring(output))
-end
+printTensorTable(output)
+
+----------------------------------------------------------------------
+
+print('******************************************************************')
+print('** updateGradInput ***********************************************')
+print('******************************************************************')
+gradInput = g:updateGradInput({ input }, output)
-torch.save('dag.t7', g)
+printTensorTable(gradInput)