small fix to make the tests pass + flake8

theano/sandbox/cuda/tests/test_abstractconv.py

@@ -46,13 +46,14 @@ class TestDnnConv2d(test_abstract_conv.BaseTestConv2d):
                            provide_shape=provide_shape, border_mode=b,
                            filter_flip=flip, target_op=GpuDnnConvGradI)
 
-class TestCorrMMConv2d(test_abstract_conv.TestConv2d):
+
+class TestCorrMMConv2d(test_abstract_conv.BaseTestConv2d):
     def setUp(self):
         super(TestCorrMMConv2d, self).setUp()
         self.shared = gpu_shared
         self.mode = mode_with_gpu.excluding('cudnn')
 
-    def test_gpucorrmm_conv(self, i, f, s, b, flip, provide_shape):
+    def tcase(self, i, f, s, b, flip, provide_shape):
         mode = self.mode
         o = self.get_output_shape(i, f, s, b)
         self.run_fwd(inputs_shape=i, filters_shape=f, subsample=s,

theano/tensor/nnet/abstract_conv.py

@@ -458,7 +458,6 @@ class BaseAbstractConv2d(Op):
                 'invalid mode {}, which must be either '
                 '"valid" or "full"'.format(mode))
 
-
         out_shape = get_conv_output_shape(img.shape, kern.shape, mode, [1, 1])
         out = numpy.zeros(out_shape, dtype=img.dtype)
         val = _valfrommode(mode)
@@ -522,16 +521,15 @@ class AbstractConv2d(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)
-            new_img[:, :, pad_h:img.shape[2]+pad_h, pad_w:img.shape[3]+pad_w] = img
+            new_img[:, :, pad_h:img.shape[2] + pad_h, pad_w:img.shape[3] + pad_w] = img
             img = new_img
         if not self.filter_flip:
             kern = kern[:, :, ::-1, ::-1]
         conv_out = self.corr2d(img, kern, mode)
-        conv_out =conv_out[:, :, ::self.subsample[0], ::self.subsample[1]]
+        conv_out = conv_out[:, :, ::self.subsample[0], ::self.subsample[1]]
 
         o[0] = node.outputs[0].type.filter(conv_out)
 
-
     def R_op(self, inputs, eval_points):
         rval = None
         if eval_points[0] is not None:
@@ -641,12 +639,11 @@ class AbstractConv2d_gradWeights(BaseAbstractConv2d):
             pad_h, pad_w = map(int, mode)
             mode = "valid"
             new_img = numpy.zeros((img.shape[0], img.shape[1],
-                                   img.shape[2] + 2 *pad_h,
+                                   img.shape[2] + 2 * pad_h,
                                    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
+            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,
@@ -664,8 +661,6 @@ class AbstractConv2d_gradWeights(BaseAbstractConv2d):
             kern = kern.transpose(1, 0, 2, 3)
         o[0] = node.outputs[0].type.filter(kern)
 
-
-
     def grad(self, inp, grads):
         bottom, top = inp[:2]
         weights, = grads
@@ -782,7 +777,7 @@ class AbstractConv2d_gradInputs(BaseAbstractConv2d):
         if self.filter_flip:
             img = img[:, :, ::-1, ::-1]
         if pad_h > 0 or pad_w > 0:
-            img = img[:, :, pad_h:img.shape[2]-pad_h, pad_w:img.shape[3]-pad_w]
+            img = img[:, :, pad_h:img.shape[2] - pad_h, pad_w:img.shape[3] - pad_w]
         o[0] = node.outputs[0].type.filter(img)
 
     def grad(self, inp, grads):

theano/tensor/nnet/tests/test_abstract_conv.py

@@ -363,32 +363,31 @@ class TestCpuConv2d(BaseTestConv2d):
                               border_mode=b,
                               filter_flip=flip)
 
-class TestDebugMode(BaseTestConv2d):
+
+class TestDebugModeConv2d(BaseTestConv2d):
     def setUp(self):
-        super(TestDnnConv2d, self).setUp()
+        super(TestDebugModeConv2d, self).setUp()
         self.provide_shape = [False]
-        self.shared = gpu_shared
 
     def tcase(self, i, f, s, b, flip, provide_shape):
         mode = "DebugMode"
         o = self.get_output_shape(i, f, s, b)
         self.run_fwd(inputs_shape=i, filters_shape=f, subsample=s,
-                     verify_grad=True, mode=mode, mode_ref=mode_without_gpu,
-                     device='cpu', provide_shape=provide_shape, border_mode=b,
+                     verify_grad=True, mode=mode,
+                     provide_shape=provide_shape, border_mode=b,
                      filter_flip=flip, target_op=None)
         self.run_gradweight(inputs_shape=i, filters_shape=f,
                             output_shape=o, subsample=s,
-                            verify_grad=True, mode=mode, mode_ref=mode_without_gpu,
-                            device='cpu', provide_shape=provide_shape, border_mode=b,
+                            verify_grad=True, mode=mode,
+                            provide_shape=provide_shape, border_mode=b,
                             filter_flip=flip, target_op=None)
         self.run_gradinput(inputs_shape=i, filters_shape=f,
                            output_shape=o, subsample=s,
-                           verify_grad=True, mode=mode, mode_ref=mode_without_gpu,
-                           device='cpu', provide_shape=provide_shape, border_mode=b,
+                           verify_grad=True, mode=mode,
+                           provide_shape=provide_shape, border_mode=b,
                            filter_flip=flip, target_op=None)
 
 
-
 class TestConvTypes(unittest.TestCase):
     def setUp(self):
         self.input = tensor.ftensor4()