[xdoctest] reformat example code with google style in No.31-35 (#56051)

python/paddle/nn/initializer/dirac.py

@@ -52,40 +52,39 @@ class Dirac(Initializer):
     Examples:
         .. code-block:: python
 
-            import paddle
-
-            #1. For kernel_size is uneven number:
-
-            attr = paddle.ParamAttr(initializer=paddle.nn.initializer.Dirac())
-            conv = paddle.nn.Conv1D(3, 2, 3, weight_attr=attr)
-            conv.weight
-            # Tensor(shape=[2, 3, 3], dtype=float32, place=CPUPlace, stop_gradient=False,
-            #       [[[0., 1., 0.],
-            #         [0., 0., 0.],
-            #         [0., 0., 0.]],
-            #
-            #        [[0., 0., 0.],
-            #         [0., 1., 0.],
-            #         [0., 0., 0.]]])
-
-            input = paddle.rand([8, 3, 10])
-            output = conv(input)
-            output == input[:, 0:2, 1:9]
-            # output.shape is [8, 2, 8], It means output is almost the same with input, 2 channels are reserved
-
-
-            #2. For kernel_size is even number:
-            attr = paddle.ParamAttr(initializer=paddle.nn.initializer.Dirac())
-            conv = paddle.nn.Conv1D(3, 2, 4, weight_attr=attr)
-            conv.weight
-            # Tensor(shape=[2, 3, 4], dtype=float32, place=CPUPlace, stop_gradient=False,
-            #       [[[0., 0., 1., 0.],
-            #         [0., 0., 0., 0.],
-            #         [0., 0., 0., 0.]],
-            #
-            #        [[0., 0., 0., 0.],
-            #         [0., 0., 1., 0.],
-            #         [0., 0., 0., 0.]]])
+            >>> import paddle
+
+            >>> # 1. For kernel_size is uneven number:
+            >>> attr = paddle.ParamAttr(initializer=paddle.nn.initializer.Dirac())
+            >>> conv = paddle.nn.Conv1D(3, 2, 3, weight_attr=attr)
+            >>> print(conv.weight)
+            Parameter containing:
+            Tensor(shape=[2, 3, 3], dtype=float32, place=CPUPlace, stop_gradient=False,
+            [[[0., 1., 0.],
+              [0., 0., 0.],
+              [0., 0., 0.]],
+             [[0., 0., 0.],
+              [0., 1., 0.],
+              [0., 0., 0.]]])
+            >>> input = paddle.rand([8, 3, 10])
+            >>> output = conv(input)
+            >>> output == input[:, 0:2, 1:9]
+            >>> print(output.shape)
+            [8, 2, 8]
+            >>> # It means output is almost the same with input, 2 channels are reserved
+
+            >>> # 2. For kernel_size is even number:
+            >>> attr = paddle.ParamAttr(initializer=paddle.nn.initializer.Dirac())
+            >>> conv = paddle.nn.Conv1D(3, 2, 4, weight_attr=attr)
+            >>> print(conv.weight)
+            Parameter containing:
+            Tensor(shape=[2, 3, 4], dtype=float32, place=CPUPlace, stop_gradient=False,
+            [[[0., 0., 1., 0.],
+              [0., 0., 0., 0.],
+              [0., 0., 0., 0.]],
+             [[0., 0., 0., 0.],
+              [0., 0., 1., 0.],
+              [0., 0., 0., 0.]]])
     """
 
     def __init__(self, groups=1, name=None):

python/paddle/nn/initializer/initializer.py

@@ -81,10 +81,10 @@ class Initializer:
         convolutions.
 
         Args:
-            var: variable for which fan_in and fan_out have to be computed
+            var: variable for which fan_in and fan_out have to be computed.
 
         Returns:
-            tuple of two integers (fan_in, fan_out)
+            tuple of two integers (fan_in, fan_out).
         """
         shape = var.shape
         if not shape or len(shape) == 0:
@@ -124,10 +124,17 @@ def calculate_gain(nonlinearity, param=None):
     Examples:
         .. code-block:: python
 
-            import paddle
-            gain = paddle.nn.initializer.calculate_gain('tanh') # 5.0 / 3
-            gain = paddle.nn.initializer.calculate_gain('leaky_relu', param=1.0) # 1.0 = math.sqrt(2.0 / (1+param^2))
-            initializer = paddle.nn.initializer.Orthogonal(gain)
+            >>> import paddle
+
+            >>> gain = paddle.nn.initializer.calculate_gain('tanh')
+            >>> print(gain)
+            1.6666666666666667
+            >>> # 5.0 / 3
+            >>> gain = paddle.nn.initializer.calculate_gain('leaky_relu', param=1.0)
+            >>> print(gain)
+            1.0
+            >>> # math.sqrt(2.0 / (1+param^2))
+            >>> initializer = paddle.nn.initializer.Orthogonal(gain)
 
     """
     if param is None:

python/paddle/nn/initializer/kaiming.py

@@ -84,7 +84,7 @@ class MSRAInitializer(Initializer):
                    should be added. Used in static graph only, default None.
 
         Returns:
-            The initialization op
+            The initialization op.
         """
         block = self._check_block(block)
 
@@ -220,14 +220,12 @@ class KaimingNormal(MSRAInitializer):
     Examples:
         .. code-block:: python
 
-            import paddle
-            import paddle.nn as nn
+            >>> import paddle
+            >>> import paddle.nn as nn
 
-            linear = nn.Linear(2,
-                               4,
-                               weight_attr=nn.initializer.KaimingNormal())
-            data = paddle.rand([30, 10, 2], dtype='float32')
-            res = linear(data)
+            >>> linear = nn.Linear(2, 4, weight_attr=nn.initializer.KaimingNormal())
+            >>> data = paddle.rand([30, 10, 2], dtype='float32')
+            >>> res = linear(data)
 
     """
 
@@ -268,14 +266,12 @@ class KaimingUniform(MSRAInitializer):
     Examples:
         .. code-block:: python
 
-            import paddle
-            import paddle.nn as nn
+            >>> import paddle
+            >>> import paddle.nn as nn
 
-            linear = nn.Linear(2,
-                               4,
-                               weight_attr=nn.initializer.KaimingUniform())
-            data = paddle.rand([30, 10, 2], dtype='float32')
-            res = linear(data)
+            >>> linear = nn.Linear(2, 4, weight_attr=nn.initializer.KaimingUniform())
+            >>> data = paddle.rand([30, 10, 2], dtype='float32')
+            >>> res = linear(data)
 
     """
 

python/paddle/nn/initializer/normal.py

@@ -50,7 +50,7 @@ class NormalInitializer(Initializer):
                    should be added. Used in static graph only, default None.
 
         Returns:
-            The initialization op
+            The initialization op.
         """
         block = self._check_block(block)
 
@@ -112,24 +112,32 @@ class Normal(NormalInitializer):
     Examples:
         .. code-block:: python
 
-            import paddle
-
-            data = paddle.ones(shape=[3, 1, 2], dtype='float32')
-            weight_attr = paddle.framework.ParamAttr(
-                name="linear_weight",
-                initializer=paddle.nn.initializer.Normal(mean=0.0, std=2.0))
-            bias_attr = paddle.framework.ParamAttr(
-                name="linear_bias",
-                initializer=paddle.nn.initializer.Normal(mean=0.0, std=2.0))
-            linear = paddle.nn.Linear(2, 2, weight_attr=weight_attr, bias_attr=bias_attr)
-            # linear.weight:  [[ 2.1973135 -2.2697184]
-            #                  [-1.9104223 -1.0541488]]
-            # linear.bias:  [ 0.7885926  -0.74719954]
-
-            res = linear(data)
-            # res:  [[[ 1.0754838 -4.071067 ]]
-            #        [[ 1.0754838 -4.071067 ]]
-            #        [[ 1.0754838 -4.071067 ]]]
+            >>> import paddle
+
+            >>> data = paddle.ones(shape=[3, 1, 2], dtype='float32')
+            >>> weight_attr = paddle.framework.ParamAttr(
+            ...     name="linear_weight",
+            ...     initializer=paddle.nn.initializer.Normal(mean=0.0, std=2.0))
+            >>> bias_attr = paddle.framework.ParamAttr(
+            ...     name="linear_bias",
+            ...     initializer=paddle.nn.initializer.Normal(mean=0.0, std=2.0))
+            >>> # doctest: +SKIP('name has been used')
+            >>> linear = paddle.nn.Linear(2, 2, weight_attr=weight_attr, bias_attr=bias_attr)
+            >>> print(linear.weight)
+            Parameter containing:
+            Tensor(shape=[2, 2], dtype=float32, place=Place(cpu), stop_gradient=False,
+            [[ 2.1973135 -2.2697184],
+             [-1.9104223 -1.0541488]])
+            >>> print(linear.bias)
+            Parameter containing:
+            Tensor(shape=[2], dtype=float32, place=Place(cpu), stop_gradient=False,
+            [ 0.7885926  -0.74719954])
+            >>> res = linear(data)
+            >>> print(res)
+            Tensor(shape=[3, 1, 2], dtype=float32, place=Place(cpu), stop_gradient=False,
+            [[[ 1.0754838 -4.071067 ]],
+             [[ 1.0754838 -4.071067 ]],
+             [[ 1.0754838 -4.071067 ]]])
     """
 
     def __init__(self, mean=0.0, std=1.0, name=None):
@@ -253,24 +261,32 @@ class TruncatedNormal(TruncatedNormalInitializer):
     Examples:
         .. code-block:: python
 
-            import paddle
-
-            data = paddle.ones(shape=[3, 1, 2], dtype='float32')
-            weight_attr = paddle.framework.ParamAttr(
-                name="linear_weight",
-                initializer=paddle.nn.initializer.TruncatedNormal(mean=0.0, std=2.0))
-            bias_attr = paddle.framework.ParamAttr(
-                name="linear_bias",
-                initializer=paddle.nn.initializer.TruncatedNormal(mean=0.0, std=2.0))
-            linear = paddle.nn.Linear(2, 2, weight_attr=weight_attr, bias_attr=bias_attr)
-            # linear.weight:  [[-1.0981836  1.4140984]
-            #                  [ 3.1390522 -2.8266568]]
-            # linear.bias:  [-2.1546738 -1.6570673]
-
-            res = linear(data)
-            # res:  [[[-0.11380529 -3.0696259 ]]
-            #        [[-0.11380529 -3.0696259 ]]
-            #        [[-0.11380529 -3.0696259 ]]
+            >>> import paddle
+
+            >>> data = paddle.ones(shape=[3, 1, 2], dtype='float32')
+            >>> weight_attr = paddle.framework.ParamAttr(
+            ...     name="linear_weight",
+            ...     initializer=paddle.nn.initializer.TruncatedNormal(mean=0.0, std=2.0))
+            >>> bias_attr = paddle.framework.ParamAttr(
+            ...     name="linear_bias",
+            ...     initializer=paddle.nn.initializer.TruncatedNormal(mean=0.0, std=2.0))
+            >>> # doctest: +SKIP('name has been used')
+            >>> linear = paddle.nn.Linear(2, 2, weight_attr=weight_attr, bias_attr=bias_attr)
+            >>> print(linear.weight)
+            Parameter containing:
+            Tensor(shape=[2, 2], dtype=float32, place=Place(cpu), stop_gradient=False,
+            [[-1.0981836  1.4140984],
+             [ 3.1390522 -2.8266568]])
+            >>> print(linear.bias)
+            Parameter containing:
+            Tensor(shape=[2], dtype=float32, place=Place(cpu), stop_gradient=False,
+            [ -2.1546738  -1.6570673])
+            >>> res = linear(data)
+            >>> print(res)
+            Tensor(shape=[3, 1, 2], dtype=float32, place=Place(cpu), stop_gradient=False,
+            [[[-0.11380529 -3.0696259 ]],
+             [[-0.11380529 -3.0696259 ]],
+             [[-0.11380529 -3.0696259 ]]])
     """
 
     def __init__(self, mean=0.0, std=1.0, name=None):

python/paddle/nn/initializer/orthogonal.py

@@ -56,14 +56,13 @@ class Orthogonal(Initializer):
     Examples:
         .. code-block:: python
 
-            import paddle
+            >>> import paddle
 
-            weight_attr = paddle.ParamAttr(initializer=paddle.nn.initializer.Orthogonal())
-            linear = paddle.nn.Linear(10, 15, weight_attr=weight_attr)
-            # linear.weight: X * X' = I
-
-            linear = paddle.nn.Linear(15, 10, weight_attr=weight_attr)
-            # linear.weight: X' * X = I
+            >>> weight_attr = paddle.ParamAttr(initializer=paddle.nn.initializer.Orthogonal())
+            >>> linear = paddle.nn.Linear(10, 15, weight_attr=weight_attr)
+            >>> # linear.weight: X * X' = I
+            >>> linear = paddle.nn.Linear(15, 10, weight_attr=weight_attr)
+            >>> # linear.weight: X' * X = I
     """
 
     def __init__(self, gain=1.0, name=None):