Refactor how the parameters are iterated over to reduce the total time of the…

theano/sandbox/cuda/tests/test_abstractconv.py

@@ -24,19 +24,14 @@ else:
 class TestDnnConv2d(test_abstract_conv.TestConv2d):
     def setUp(self):
         super(TestDnnConv2d, self).setUp()
+        # provide_shape is not used by the CuDNN impementation
+        self.provide_shape = [False]
         self.shared = gpu_shared
 
-    def test_dnn_conv(self):
+    def tcase(self, i, f, s, b, flip, provide_shape):
         if not dnn_available():
             raise SkipTest(cuda.dnn.dnn_available.msg)
         mode = mode_with_gpu
-        # provide_shape is not used by the CuDNN impementation
-        provide_shape = False
-        for (i, f), s, b, flip in itertools.product(
-                zip(self.inputs_shapes, self.filters_shapes),
-                self.subsamples,
-                self.border_modes,
-                self.filter_flip):
         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,
@@ -58,16 +53,10 @@ class TestCorrMMConv2d(test_abstract_conv.TestConv2d):
     def setUp(self):
         super(TestCorrMMConv2d, self).setUp()
         self.shared = gpu_shared
+        self.mode = mode_with_gpu.excluding('cudnn')
 
-    def test_gpucorrmm_conv(self):
-        mode = mode_with_gpu.excluding('cudnn')
-        for (i, f), s, b, flip, provide_shape in itertools.product(
-                zip(self.inputs_shapes, self.filters_shapes),
-                self.subsamples,
-                self.border_modes,
-                self.filter_flip,
-                [False, True]):
-
+    def test_gpucorrmm_conv(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,
                      verify_grad=True, mode=mode,

theano/sandbox/gpuarray/tests/test_abstractconv.py

@@ -15,19 +15,14 @@ class TestDnnConv2d(test_abstract_conv.TestConv2d):
     def setUp(self):
         super(TestDnnConv2d, self).setUp()
         self.shared = gpuarray_shared_constructor
+        # provide_shape is not used by the CuDNN impementation
+        self.provide_shape = [False]
 
-    def test_dnn_conv(self):
+    def tcase(self, i, f, s, b, flip, provide_shape):
         if not dnn_available(test_ctx_name):
             raise SkipTest(dnn_available.msg)
         mode = mode_with_gpu
-        # provide_shape is not used by the CuDNN impementation
-        provide_shape = False
 
-        for (i, f), s, b, flip in itertools.product(
-                zip(self.inputs_shapes, self.filters_shapes),
-                self.subsamples,
-                self.border_modes,
-                self.filter_flip):
         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,

theano/tensor/nnet/tests/test_abstract_conv.py

@@ -67,6 +67,7 @@ class TestConv2d(unittest.TestCase):
         self.subsamples = [(1, 1), (2, 2), (2, 4)]
         self.border_modes = ["valid", "full", (0, 0), (1, 1), (5, 5), (5, 2)]
         self.filter_flip = [True, False]
+        self.provide_shape = [True, False]
         self.shared = theano.compile.shared
 
     def get_output_shape(self, inputs_shape, filters_shape, subsample,
@@ -237,16 +238,24 @@ class TestConv2d(unittest.TestCase):
                             [filters_val, output_val],
                             mode=mode, eps=1)
 
+    def test_all(self):
+        ds = [0, 0]
+        db = (0, 0)
+        dflip = True in self.filter_flip
+        dprovide_shape = True in self.provide_shapes
+        for (i, f) in zip(self.inputs_shapes, self.filters_shapes):
+            for provide_shape in self.provide_shape:
+                self.tcase(i, f, ds, db, dflip, provide_shape)
+            for s in self.subsamples:
+                self.tcase(i, f, s, db, dflip, dprovide_shape)
+            for b in self.border_modes:
+                self.tcase(i, f, ds, b, dflip, dprovide_shape)
+            for flip in self.filter_flip:
+                self.tcase(i, f, ds, db, flip, dprovide_shape)
 
-class TestCorrConv2d(TestConv2d):
-    def test_corrmm_conv(self):
-        for (i, f), s, b, flip, provide_shape in itertools.product(
-                zip(self.inputs_shapes, self.filters_shapes),
-                self.subsamples,
-                self.border_modes,
-                self.filter_flip,
-                [False, True]):
 
+class TestCorrConv2d(TestConv2d):
+    def tcase(self, i, f, s, b, flip, provide_shape):
         o = self.get_output_shape(i, f, s, b)
         self.run_fwd(inputs_shape=i, filters_shape=f, subsample=s,
                      verify_grad=True, provide_shape=provide_shape,
@@ -262,15 +271,11 @@ class TestCorrConv2d(TestConv2d):
 
 
 class TestCpuConv2d(TestConv2d):
-    def test_cpu_conv(self):
-        mode = theano.compile.mode.get_default_mode().excluding('conv_gemm')
-        for (i, f), s, b, flip, provide_shape in itertools.product(
-                zip(self.inputs_shapes, self.filters_shapes),
-                self.subsamples,
-                self.border_modes,
-                self.filter_flip,
-                [False, True]):
-
+    def setUp(self):
+        super(TestCpuConv2d, self).setUp()
+        self.mode = theano.compile.mode.get_default_mode().excluding('conv_gemm')
+    def tcase(self, i, f, s, b, flip, provide_shape):
+        mode = self.mode
         o = self.get_output_shape(i, f, s, b)
         fwd_OK = True
         gradweight_OK = True
@@ -296,9 +301,9 @@ class TestCpuConv2d(TestConv2d):
 
         if fwd_OK:
             self.run_fwd(inputs_shape=i, filters_shape=f, subsample=s,
-                             verify_grad=False, mode=mode,
-                             provide_shape=provide_shape, border_mode=b,
-                             filter_flip=flip, target_op=ConvOp)
+                         verify_grad=(gradweights_ok and gradinput_ok),
+                         mode=mode, provide_shape=provide_shape,
+                         border_mode=b, filter_flip=flip, target_op=ConvOp)
         else:
             self.assertRaises(NotImplementedError,
                               self.run_fwd,