Merge branch 'nn_fast_reset'

12 files changed, 128 insertions, 78 deletions

diff --git a/Euclidean.lua b/Euclidean.lua
index 808b7ab..c0dd99c 100644
--- a/Euclidean.lua
+++ b/Euclidean.lua
@@ -20,11 +20,14 @@ function Euclidean:reset(stdv)
    else
       stdv = 1./math.sqrt(self.weight:size(1))
    end
-
-   for i=1,self.weight:size(2) do
-      self.weight:select(2, i):apply(function()
-                                        return torch.uniform(-stdv, stdv)
-                                     end)
+   if nn.oldSeed then
+      for i=1,self.weight:size(2) do
+         self.weight:select(2, i):apply(function()
+            return torch.uniform(-stdv, stdv)
+         end)
+      end
+   else
+      self.weight:uniform(-stdv, stdv)
    end
 end
 
diff --git a/Linear.lua b/Linear.lua
index 953af78..cc6da4e 100644
--- a/Linear.lua
+++ b/Linear.lua
@@ -17,14 +17,16 @@ function Linear:reset(stdv)
    else
       stdv = 1./math.sqrt(self.weight:size(2))
    end
-
-   -- we do this so the initialization is exactly
-   -- the same than in previous torch versions
-   for i=1,self.weight:size(1) do
-      self.weight:select(1, i):apply(function()
-                                        return torch.uniform(-stdv, stdv)
-                                     end)
-      self.bias[i] = torch.uniform(-stdv, stdv)
+   if nn.oldSeed then
+      for i=1,self.weight:size(1) do
+         self.weight:select(1, i):apply(function()
+            return torch.uniform(-stdv, stdv)
+         end)
+         self.bias[i] = torch.uniform(-stdv, stdv)
+      end
+   else
+      self.weight:uniform(-stdv, stdv)
+      self.bias:uniform(-stdv, stdv)
    end
 end
 
diff --git a/LookupTable.lua b/LookupTable.lua
index 7dfda7a..671783f 100644
--- a/LookupTable.lua
+++ b/LookupTable.lua
@@ -29,9 +29,13 @@ end
 
 function LookupTable:reset(stdv)
    stdv = stdv or 1
-   self.weight:apply(function()
-                        return torch.normal(0, stdv)
-                     end)
+   if nn.oldSeed then
+      self.weight:apply(function()
+         return torch.normal(0, stdv)
+      end)
+   else
+      self.weight:normal(0, stdv)
+   end
 end
 
 function LookupTable:updateOutput(input)
diff --git a/SparseLinear.lua b/SparseLinear.lua
index ec8845e..f1a2be5 100644
--- a/SparseLinear.lua
+++ b/SparseLinear.lua
@@ -22,14 +22,16 @@ function SparseLinear:reset(stdv)
    else
       stdv = 1./math.sqrt(self.weight:size(1))
    end
-
-   -- we do this so the initialization is exactly
-   -- the same than in previous torch versions
-   for i=1,self.weight:size(1) do
-      self.weight:select(1, i):apply(function()
-                                        return torch.uniform(-stdv, stdv)
-                                     end)
-      self.bias[i] = torch.uniform(-stdv, stdv) * 0.000001
+   if nn.oldSeed then
+      for i=1,self.weight:size(1) do
+         self.weight:select(1, i):apply(function()
+            return torch.uniform(-stdv, stdv)
+         end)
+         self.bias[i] = torch.uniform(-stdv, stdv) * 0.000001
+      end
+   else
+      self.weight:uniform(-stdv, stdv)
+      self.bias:uniform(-stdv, stdv):mul(0.000001)
    end
 end
 
diff --git a/SpatialConvolution.lua b/SpatialConvolution.lua
index 38d2737..e5e8fc9 100644
--- a/SpatialConvolution.lua
+++ b/SpatialConvolution.lua
@@ -27,12 +27,17 @@ function SpatialConvolution:reset(stdv)
    else
       stdv = 1/math.sqrt(self.kW*self.kH*self.nInputPlane)
    end
-   self.weight:apply(function()
-                        return torch.uniform(-stdv, stdv)
-                     end)
-   self.bias:apply(function()
-                      return torch.uniform(-stdv, stdv)
-                   end)   
+   if nn.oldSeed then
+      self.weight:apply(function()
+         return torch.uniform(-stdv, stdv)
+      end)
+      self.bias:apply(function()
+         return torch.uniform(-stdv, stdv)
+      end) 
+   else
+      self.weight:uniform(-stdv, stdv)
+      self.bias:uniform(-stdv, stdv)
+   end
 end
 
 function SpatialConvolution:updateOutput(input)
diff --git a/SpatialConvolutionMM.lua b/SpatialConvolutionMM.lua
index aa358e5..cc6a718 100644
--- a/SpatialConvolutionMM.lua
+++ b/SpatialConvolutionMM.lua
@@ -25,12 +25,17 @@ function SpatialConvolutionMM:reset(stdv)
    else
       stdv = 1/math.sqrt(self.kW*self.kH*self.nInputPlane)
    end
-   self.weight:apply(function()
-                        return torch.uniform(-stdv, stdv)
-                     end)
-   self.bias:apply(function()
-                      return torch.uniform(-stdv, stdv)
-                   end)   
+   if nn.oldSeed then
+      self.weight:apply(function()
+         return torch.uniform(-stdv, stdv)
+      end)
+      self.bias:apply(function()
+         return torch.uniform(-stdv, stdv)
+      end)  
+   else
+      self.weight:uniform(-stdv, stdv)
+      self.bias:uniform(-stdv, stdv)
+   end
 end
 
 function SpatialConvolutionMM:updateOutput(input)
diff --git a/SpatialConvolutionMap.lua b/SpatialConvolutionMap.lua
index 4f6609e..f99668b 100644
--- a/SpatialConvolutionMap.lua
+++ b/SpatialConvolutionMap.lua
@@ -110,22 +110,31 @@ end
 function SpatialConvolutionMap:reset(stdv)
    if stdv then
       stdv = stdv * math.sqrt(3)
-      self.weight:apply(function()
-			   return torch.uniform(-stdv, stdv)
-			end)
-      self.bias:apply(function()
-			 return torch.uniform(-stdv, stdv)
-		      end)
+      if nn.oldSeed then
+         self.weight:apply(function()
+            return torch.uniform(-stdv, stdv)
+         end)
+         self.bias:apply(function()
+            return torch.uniform(-stdv, stdv)
+         end)
+      else
+         self.weight:uniform(-stdv, stdv)
+         self.bias:uniform(-stdv, stdv)
+      end
    else
       local ninp = torch.Tensor(self.nOutputPlane):zero()
       for i=1,self.connTable:size(1) do ninp[self.connTable[i][2]] =  ninp[self.connTable[i][2]]+1 end
       for k=1,self.connTable:size(1) do
-	 stdv = 1/math.sqrt(self.kW*self.kH*ninp[self.connTable[k][2]])
-	 self.weight:select(1,k):apply(function() return torch.uniform(-stdv,stdv) end)
+         stdv = 1/math.sqrt(self.kW*self.kH*ninp[self.connTable[k][2]])
+         if nn.oldSeed then
+            self.weight:select(1,k):apply(function() return torch.uniform(-stdv,stdv) end)
+         else
+            self.weight:select(1,k):uniform(-stdv,stdv)
+         end
       end
       for k=1,self.bias:size(1) do
-	 stdv = 1/math.sqrt(self.kW*self.kH*ninp[k])
-	 self.bias[k] = torch.uniform(-stdv,stdv)
+         stdv = 1/math.sqrt(self.kW*self.kH*ninp[k])
+         self.bias[k] = torch.uniform(-stdv,stdv)
       end
    end
 end
diff --git a/SpatialSubSampling.lua b/SpatialSubSampling.lua
index 48b32b9..19aa2f5 100644
--- a/SpatialSubSampling.lua
+++ b/SpatialSubSampling.lua
@@ -26,12 +26,17 @@ function SpatialSubSampling:reset(stdv)
    else
       stdv = 1/math.sqrt(self.kW*self.kH)
    end
-   self.weight:apply(function()
-                        return torch.uniform(-stdv, stdv)
-                     end)
-   self.bias:apply(function()
-                      return torch.uniform(-stdv, stdv)
-                   end)   
+   if nn.oldSeed then
+      self.weight:apply(function()
+         return torch.uniform(-stdv, stdv)
+      end)
+      self.bias:apply(function()
+         return torch.uniform(-stdv, stdv)
+      end) 
+   else
+      self.weight:uniform(-stdv, stdv)
+      self.bias:uniform(-stdv, stdv)
+   end
 end
 
 function SpatialSubSampling:updateOutput(input)
diff --git a/TemporalConvolution.lua b/TemporalConvolution.lua
index a3aaa7f..b71e4c7 100644
--- a/TemporalConvolution.lua
+++ b/TemporalConvolution.lua
@@ -24,12 +24,17 @@ function TemporalConvolution:reset(stdv)
    else
       stdv = 1/math.sqrt(self.kW*self.inputFrameSize)
    end
-   self.weight:apply(function()
-                        return torch.uniform(-stdv, stdv)
-                     end)
-   self.bias:apply(function()
-                      return torch.uniform(-stdv, stdv)
-                   end)   
+   if nn.oldSeed then
+      self.weight:apply(function()
+         return torch.uniform(-stdv, stdv)
+      end)
+      self.bias:apply(function()
+         return torch.uniform(-stdv, stdv)
+      end)   
+   else
+      self.weight:uniform(-stdv, stdv)
+      self.bias:uniform(-stdv, stdv)
+   end
 end
 
 function TemporalConvolution:updateOutput(input)
diff --git a/TemporalSubSampling.lua b/TemporalSubSampling.lua
index 3d06f6e..29df8b1 100644
--- a/TemporalSubSampling.lua
+++ b/TemporalSubSampling.lua
@@ -23,14 +23,17 @@ function TemporalSubSampling:reset(stdv)
    else
       stdv = 1/math.sqrt(self.kW)
    end
-
-   self.weight:apply(function()
-                        return torch.uniform(-stdv, stdv)
-                     end)
-
-   self.bias:apply(function()
-                      return torch.uniform(-stdv, stdv)
-                   end)   
+   if nn.oldSeed then
+      self.weight:apply(function()
+         return torch.uniform(-stdv, stdv)
+      end)
+      self.bias:apply(function()
+         return torch.uniform(-stdv, stdv)
+      end)
+   else
+      self.weight:uniform(-stdv, stdv)
+      self.bias:uniform(-stdv, stdv)
+   end
 end
 
 function TemporalSubSampling:updateOutput(input)
diff --git a/VolumetricConvolution.lua b/VolumetricConvolution.lua
index 4262199..4dec9e3 100644
--- a/VolumetricConvolution.lua
+++ b/VolumetricConvolution.lua
@@ -30,12 +30,17 @@ function VolumetricConvolution:reset(stdv)
    else
       stdv = 1/math.sqrt(self.kT*self.kW*self.kH*self.nInputPlane)
    end
-   self.weight:apply(function()
-                        return torch.uniform(-stdv, stdv)
-                     end)
-   self.bias:apply(function()
-                      return torch.uniform(-stdv, stdv)
-                   end)   
+   if nn.oldSeed then
+      self.weight:apply(function()
+         return torch.uniform(-stdv, stdv)
+      end)
+      self.bias:apply(function()
+         return torch.uniform(-stdv, stdv)
+      end)
+   else
+      self.weight:uniform(-stdv, stdv)
+      self.bias:uniform(-stdv, stdv)
+   end
 end
 
 function VolumetricConvolution:updateOutput(input)
diff --git a/WeightedEuclidean.lua b/WeightedEuclidean.lua
index 5337eec..c4a1dbc 100644
--- a/WeightedEuclidean.lua
+++ b/WeightedEuclidean.lua
@@ -30,13 +30,15 @@ function WeightedEuclidean:reset(stdv)
    else
       stdv = 1./math.sqrt(self.templates:size(1))
    end
-
-   for i=1,self.templates:size(2) do
-      self.templates:select(2, i):apply(function()
-                                        return torch.uniform(-stdv, stdv)
-                                     end)
+   if nn.oldSeed then
+      for i=1,self.templates:size(2) do
+         self.templates:select(2, i):apply(function()
+            return torch.uniform(-stdv, stdv)
+         end)
+      end
+   else
+      self.templates:uniform(-stdv, stdv)
    end
-
    self.diagCov:fill(1)
 end