Changed comments to docstrings where necessary in theano/tensor/nnet/tests

theano/tensor/nnet/tests/test_abstract_conv.py

@@ -1097,12 +1097,16 @@ class TestBilinearUpsampling(unittest.TestCase):
         compile_mode = compile_mode.excluding('AbstractConvCheck')
 
     def numerical_kernel_1D(self, ratio):
-        # Gets numerical 1D kernel for bilinear upsampling
+        """
+        Gets numerical 1D kernel for bilinear upsampling
+        """
         return np.array(list(range(1, ratio + 1)) +
                         list(range(ratio - 1, 0, -1)))
 
     def numerical_kernel_2D(self, ratio):
-        # Gets numerical 2D kernel for bilinear upsampling
+        """
+        Gets numerical 2D kernel for bilinear upsampling
+        """
         return np.array([i * j for i in self.numerical_kernel_1D(ratio) for j
                          in self.numerical_kernel_1D(ratio)]).\
             reshape(2 * ratio - 1, 2 * ratio - 1)
@@ -1157,50 +1161,52 @@ class TestBilinearUpsampling(unittest.TestCase):
             utt.assert_allclose(kernel_1D, f_ten_norm(ratio))
 
     def numerical_upsampling_multiplier(self, ratio):
-        # Compute upsampling multiplier
-        #
-        # This method computes the multipliers of an array
-        # that will be upsampled using bilinear interpolation.
-        #
-        # Parameters
-        # ----------
-        # ratio: int
-        #     the ratio by which the array will be upsampled.
-        #
-        # Returns
-        # -------
-        # 1D numpy array
-        #     The multiplers that can be used in bilinear interpolation
-        #     to upsample an array.
-        #
-        # int
-        #     The size of the multipliers array
-
+        """
+        Compute upsampling multiplier
+
+        This method computes the multipliers of an array
+        that will be upsampled using bilinear interpolation.
+
+        Parameters
+        ----------
+        ratio: int
+            the ratio by which the array will be upsampled.
+
+        Returns
+        -------
+        1D numpy array
+            The multiplers that can be used in bilinear interpolation
+            to upsample an array.
+
+        int
+            The size of the multipliers array
+        """
         kern = np.arange(ratio + 1)
         return kern, kern.shape[0]
 
     def get_upsampled_twobytwo_mat(self, two_by_two, ratio):
-        # Upsample 4D array with two rows and two columns
-        #
-        # This method gets a 4D numpy array with two rows and two columns
-        # and computes its upsampled array by using bilinear interpolation
-        #
-        # Parameters
-        # ----------
-        # two_by_two: numpy 4D array
-        #     The array that will be upsampled by bilinear interpolation.
-        #     Array is of shape (batch size, num channels, 2, 2)
-        #
-        # ratio: int
-        #     The ratio by which two_by_two's last
-        #     two dimensions (row and col) will be upsampled.
-        #
-        # Returns
-        # -------
-        # 4D numpy array
-        #     The array upsampled by using bilinear interpolation. Array
-        #     is of shape (batch size, num channels, 2*ratio, 2*ratio).
-
+        """
+        Upsample 4D array with two rows and two columns
+
+        This method gets a 4D numpy array with two rows and two columns
+        and computes its upsampled array by using bilinear interpolation
+
+        Parameters
+        ----------
+        two_by_two: numpy 4D array
+            The array that will be upsampled by bilinear interpolation.
+            Array is of shape (batch size, num channels, 2, 2)
+
+        ratio: int
+            The ratio by which two_by_two's last
+            two dimensions (row and col) will be upsampled.
+
+        Returns
+        -------
+        4D numpy array
+            The array upsampled by using bilinear interpolation. Array
+            is of shape (batch size, num channels, 2*ratio, 2*ratio).
+        """
         kern, shp = self.numerical_upsampling_multiplier(ratio)
         up_1D = two_by_two[:, :, :, :1] * kern[::-1] + \
             two_by_two[:, :, :, 1:] * kern

theano/tensor/nnet/tests/test_blocksparse.py

@@ -87,8 +87,9 @@ class BlockSparse_Gemv_and_Outer(utt.InferShapeTester):
 
     @staticmethod
     def gemv_numpy2(o, W, h, iIdx, oIdx):
-        # Other implementation
-
+        """
+        Other implementation
+        """
         from numpy import ix_
         for b in range(o.shape[0]):
             w = W[ix_(iIdx[b], oIdx[b])].swapaxes(1, 2)
@@ -98,8 +99,9 @@ class BlockSparse_Gemv_and_Outer(utt.InferShapeTester):
 
     @staticmethod
     def gemv_numpy3(o, W, h, iIdx, oIdx):
-        # Other implementation
-
+        """
+        Other implementation
+        """
         from numpy import ix_
         for b in range(o.shape[0]):
             w = W[ix_(iIdx[b], oIdx[b])]

theano/tensor/nnet/tests/test_conv.py

@@ -36,16 +36,18 @@ class TestConv2D(utt.InferShapeTester):
                  input=None, filters=None,
                  unroll_batch=None, unroll_kern=None, unroll_patch=None,
                  verify_grad=True, should_raise=False):
-        # :param image_shape: The constant shape info passed to conv2d.
-        # :param filter_shape: The constant shape info passed to conv2d.
-        #
-        # :param N_image_shape: None(default to image_shape) or tuple of
-        #                       4 elements with the shape of the input image
-        #
-        # :param N_filter_shape: None(default to filter_shape) or tuple
-        #                        of 4 elements with the shape of the
-        #                        input filter
+        """
+        :param image_shape: The constant shape info passed to conv2d.
+        :param filter_shape: The constant shape info passed to conv2d.
+
+        :param N_image_shape: None(default to image_shape) or tuple of
+                              4 elements with the shape of the input image
+
+        :param N_filter_shape: None(default to filter_shape) or tuple
+                               of 4 elements with the shape of the
+                               input filter
 
+        """
         if N_image_shape is None:
             N_image_shape = [T.get_scalar_constant_value(
                 T.as_tensor_variable(x)) for x in image_shape]

theano/tensor/nnet/tests/test_corr.py

@@ -34,9 +34,10 @@ class TestCorr2D(utt.InferShapeTester):
                  border_mode='valid', subsample=(1, 1),
                  input=None, filters=None, verify_grad=True,
                  non_contiguous=False, filter_dilation=(1, 1)):
-        # :param image_shape: The constant shape info passed to corrMM.
-        # :param filter_shape: The constant shape info passed to corrMM.
-
+        """
+        :param image_shape: The constant shape info passed to corrMM.
+        :param filter_shape: The constant shape info passed to corrMM.
+        """
         if not theano.config.cxx:
             raise SkipTest("Need cxx to test conv2d")
         N_image_shape = [T.get_scalar_constant_value(T.as_tensor_variable(x))

theano/tensor/nnet/tests/test_corr3d.py

@@ -36,9 +36,10 @@ class TestCorr3D(utt.InferShapeTester):
                  border_mode='valid', subsample=(1, 1, 1),
                  input=None, filters=None, verify_grad=True,
                  non_contiguous=False, filter_dilation=(1, 1, 1)):
-        # :param image_shape: The constant shape info passed to corr3dMM.
-        # :param filter_shape: The constant shape info passed to corr3dMM.
-
+        """
+        :param image_shape: The constant shape info passed to corr3dMM.
+        :param filter_shape: The constant shape info passed to corr3dMM.
+        """
         if not theano.config.cxx:
             raise SkipTest("Need cxx for this test")
 

theano/tensor/nnet/tests/test_ctc.py

@@ -77,9 +77,12 @@ def setup_grad_case():
 
 
 class TestCTC(unittest.TestCase):
-    # Test Baidu CTC wrapper implementation.
-    # Expected values for costs and gradients are obtained through an external
-    # C implementation, that uses the library directly.
+    """
+    Test Baidu CTC wrapper implementation.
+
+    Expected values for costs and gradients are obtained through an external
+    C implementation, that uses the library directly.
+    """
 
     def setUp(self):
         if theano.config.mode == "FAST_COMPILE" or theano.config.cxx == "":
@@ -108,8 +111,9 @@ class TestCTC(unittest.TestCase):
         self.check_grads_disabled(t_activations, t_labels, t_activation_times)
 
     def check_grads_disabled(self, activations, labels, input_length):
-        # Check if optimization to disable gradients is working
-
+        """
+        Check if optimization to disable gradients is working
+        """
         ctc_cost = ctc(activations, labels, input_length)
         ctc_function = theano.function([], [ctc_cost])
         for node in ctc_function.maker.fgraph.apply_nodes:

theano/tensor/nnet/tests/test_sigm.py

@@ -87,14 +87,15 @@ class T_softplus(unittest.TestCase):
 class T_sigmoid_opts(unittest.TestCase):
 
     def get_mode(self, excluding=None):
-        # Return appropriate mode for the tests.
-        #
-        # :param excluding: List of optimizations to exclude.
-        #
-        # :return: The current default mode unless the `config.mode` option is
-        # set to 'FAST_COMPILE' (in which case it is replaced by the 'FAST_RUN'
-        # mode), without the optimizations specified in `excluding`.
+        """
+        Return appropriate mode for the tests.
+
+        :param excluding: List of optimizations to exclude.
 
+        :return: The current default mode unless the `config.mode` option is
+        set to 'FAST_COMPILE' (in which case it is replaced by the 'FAST_RUN'
+        mode), without the optimizations specified in `excluding`.
+        """
         if excluding is None:
             excluding = []
         m = theano.config.mode
@@ -447,7 +448,9 @@ class T_softplus_opts(unittest.TestCase):
 
 
 class T_sigmoid_utils(unittest.TestCase):
-    # Test utility functions found in 'sigm.py'.
+    """
+    Test utility functions found in 'sigm.py'.
+    """
 
     def test_compute_mul(self):
         x, y, z = tensor.vectors('x', 'y', 'z')