X-Git-Url: https://www.fleuret.org/cgi-bin/gitweb/gitweb.cgi?p=dagnn.git;a=blobdiff_plain;f=test-dagnn.lua;h=a41d8802931ec51f500a42ba4e2608540addf340;hp=262ea6fe3111830ab1f8270118b608725e124881;hb=da6186a657b7563841416c42336e52937b76d67f;hpb=452781856eafd237579e5c90b6e345354df91b42

diff --git a/test-dagnn.lua b/test-dagnn.lua
index 262ea6f..a41d880 100755
--- a/test-dagnn.lua
+++ b/test-dagnn.lua
@@ -1,43 +1,118 @@
 #!/usr/bin/env luajit
 
+--[[
+
+   Copyright (c) 2016 Idiap Research Institute, http://www.idiap.ch/
+   Written by Francois Fleuret <francois.fleuret@idiap.ch>
+
+   This file is free software: you can redistribute it and/or modify
+   it under the terms of the GNU General Public License version 3 as
+   published by the Free Software Foundation.
+
+   It is distributed in the hope that it will be useful, but WITHOUT
+   ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
+   or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public
+   License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this file.  If not, see <http://www.gnu.org/licenses/>.
+
+]]--
+
 require 'torch'
 require 'nn'
-
 require 'dagnn'
 
-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)
+-- torch.setnumthreads(params.nbThreads)
+torch.setdefaulttensortype('torch.DoubleTensor')
+torch.manualSeed(2)
 
---[[
+function checkGrad(model, criterion, input, target)
+   local params, gradParams = model:getParameters()
 
-   a -----> b ---> c ----> e ---
-             \           /
-              \--> d ---/
-                    \
-                     \---> f ---
-]]--
+   local epsilon = 1e-5
 
-g = nn.DAG:new()
+   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()
 
-g:setInput(a)
-g:setOutput({ e, f })
+   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
+
+      if num == ana then
+         err = 0
+      else
+         err = torch.abs(num - ana) / torch.abs(num)
+      end
+
+      print(
+         'CHECK '
+            .. 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
+
+--               +- Linear(10, 10) -> ReLU ---> d --+
+--              /                              /     \
+--             /                              /       \
+--  --> a --> b -----------> c --------------+         e -->
+--                            \                       /
+--                             \                     /
+--                              +-- Mul(-1) --------+
+
+model = nn.DAG()
+
+a = nn.Linear(50, 10)
+b = nn.ReLU()
+c = nn.Linear(10, 15)
+d = nn.CMulTable()
+e = nn.CAddTable()
 
-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:addEdge(a, b)
+model:addEdge(b, nn.Linear(10, 15), nn.ReLU(), d)
+model:addEdge(d, e)
+model:addEdge(b, c)
+model:addEdge(c, d)
+model:addEdge(c, nn.Mul(-1), e)
 
-g:print()
+model:setInput(a)
+model:setOutput(e)
 
-input = torch.Tensor(3, 10):uniform()
+local input = torch.Tensor(30, 50):uniform()
+local output = model:updateOutput(input):clone()
 
-output = g:updateOutput(input)
+output:uniform()
 
-print(output[1])
-print(output[2])
+checkGrad(model, nn.MSECriterion(), input, output)