Improve tests to check grad for all three passes

theano/tensor/nnet/tests/test_abstract_conv.py

@@ -1743,6 +1743,10 @@ class Grouped_conv_noOptim(unittest.TestCase):
 
         utt.assert_allclose(grouped_output, normal_concat_output)
 
+        utt.verify_grad(grouped_abstractconv_func,
+                        [img, kern],
+                        mode=self.mode)
+
     def test_gradweights(self):
         img = np.random.random(self.img_shape).astype('float32')
         top = np.random.random(self.top_shape).astype('float32')
@@ -1775,6 +1779,13 @@ class Grouped_conv_noOptim(unittest.TestCase):
 
         utt.assert_allclose(grouped_output, normal_concat_output)
 
+        def abstract_conv_gradweight(inputs_val, output_val):
+            return grouped_abstractconvgrad_func(inputs_val, output_val, self.kern_shape[-2:])
+
+            utt.verify_grad(abstract_conv_gradweight,
+                            [img, top],
+                            mode=self.mode, eps=1)
+
     def test_gradinputs(self):
         kern = np.random.random(self.kern_shape).astype('float32')
         top = np.random.random(self.top_shape).astype('float32')
@@ -1807,3 +1818,10 @@ class Grouped_conv_noOptim(unittest.TestCase):
         normal_concat_output = np.concatenate(normal_concat_output, axis=1)
 
         utt.assert_allclose(grouped_output, normal_concat_output)
+
+        def abstract_conv_gradinputs(filters_val, output_val):
+            return grouped_abstractconvgrad_func(filters_val, output_val, self.img_shape[2:])
+
+            utt.verify_grad(abstract_conv_gradinputs,
+                            [kern, top],
+                            mode=self.mode, eps=1)