Fix border_mode

theano/tensor/nnet/abstract_conv.py

@@ -510,8 +510,9 @@ class AbstractConv2d(BaseAbstractConv2d):
         o, = out_
         mode = self.border_mode
 
-        ### Pad
-        if mode == "half":
+        if mode == "full":
+           mode = (kern.shape[2] - 1, kern.shape[3] - 1)
+        elif mode == "half":
             mode = (kern.shape[2] // 2, kern.shape[3] // 2)
         if isinstance(mode, tuple):
             pad_h, pad_w = map(int, mode)
@@ -521,14 +522,11 @@ class AbstractConv2d(BaseAbstractConv2d):
                                    img.shape[3] + 2 * pad_w), dtype=img.dtype)
             new_img[:, :, pad_h:img.shape[2]+pad_h, pad_w:img.shape[3]+pad_w] = img
             img = new_img
-        ### Filter flip
         if not self.filter_flip:
             kern = kern[:, :, ::-1, ::-1]
-
         conv_out = self.corr2d(img, kern, mode)
-
-        ### Subsample
         conv_out =conv_out[:, :, ::self.subsample[0], ::self.subsample[1]]
+
         o[0] = conv_out
 
 
@@ -630,7 +628,9 @@ class AbstractConv2d_gradWeights(BaseAbstractConv2d):
         o, = out_
 
         mode = self.border_mode
-        if mode == "half":
+        if mode == "full":
+           mode = (shape[0] - 1, shape[1] - 1)
+        elif mode == "half":
             mode = (shape[0] // 2, shape[1] // 2)
         if isinstance(mode, tuple):
             pad_h, pad_w = map(int, mode)
@@ -638,9 +638,11 @@ class AbstractConv2d_gradWeights(BaseAbstractConv2d):
             new_img = numpy.zeros((img.shape[0], img.shape[1],
                                    img.shape[2] + 2 *pad_h,
                                    img.shape[3] + 2 * pad_w), dtype=img.dtype)
+            #import pdb; pdb.set_trace()
             new_img[:, :, pad_h:img.shape[2]+pad_h, pad_w:img.shape[3]+pad_w] = img
             img = new_img
 
+
         if self.subsample[0] > 1 or self.subsample[1] > 1:
             new_shape = (topgrad.shape[0], topgrad.shape[1],
                          img.shape[2] - shape[0] + 1,
@@ -752,9 +754,14 @@ class AbstractConv2d_gradInputs(BaseAbstractConv2d):
 
         mode = self.border_mode
         pad_h, pad_w = 0, 0
-        if isinstance(mode, tuple):
-            mode = "valid"
+
+        if mode == "full":
+            pad_h, pad_w = (kern.shape[2] - 1, kern.shape[3] - 1)
+        elif mode == "half":
+            pad_h, pad_w =  (kern.shape[2] // 2, kern.shape[3] // 2)
+        elif isinstance(mode, tuple):
             pad_h, pad_w = map(int, self.border_mode)
+        mode = "valid"
 
         if self.subsample[0] > 1 or self.subsample[1] > 1:
             new_shape = (topgrad.shape[0], topgrad.shape[1],
@@ -769,8 +776,7 @@ class AbstractConv2d_gradInputs(BaseAbstractConv2d):
         img = self.corr2d(topgrad, kern, mode="full")
         if self.filter_flip:
             img = img[:, :, ::-1, ::-1]
-        if isinstance(self.border_mode, tuple):
-            pad_h, pad_w = map(int, self.border_mode)
+        if pad_h > 0 or pad_w > 0:
             img = img[:, :, pad_h:img.shape[2]-pad_h, pad_w:img.shape[2]-pad_w]
         o[0] = img