Fix some english doc api, shenze tools group (#20300)

* test=document_fix
Fix english doc api, invloves the op of greater_equal,greater_than,less_equal,not_equal,
rank,rsqrt,diag,linspace,reduce_all,reduce_any,sign,where,zeros_like,unique_with_counts.

* Fix some format problem in the op of sign and greather_than.
test=develop
test=document_fix

* Fix the example of zeros_like, and update api.spec
test=develop
test=document_fix

paddle/fluid/API.spec

@@ -167,8 +167,8 @@ paddle.fluid.layers.reduce_mean (ArgSpec(args=['input', 'dim', 'keep_dim', 'name
 paddle.fluid.layers.reduce_max (ArgSpec(args=['input', 'dim', 'keep_dim', 'name'], varargs=None, keywords=None, defaults=(None, False, None)), ('document', '10023caec4d7f78c3b901f023a1feaa7'))
 paddle.fluid.layers.reduce_min (ArgSpec(args=['input', 'dim', 'keep_dim', 'name'], varargs=None, keywords=None, defaults=(None, False, None)), ('document', '1a1c91625ce3c32646f69ca10d4d1da7'))
 paddle.fluid.layers.reduce_prod (ArgSpec(args=['input', 'dim', 'keep_dim', 'name'], varargs=None, keywords=None, defaults=(None, False, None)), ('document', 'b386471f0476c80c61d8c8672278063d'))
-paddle.fluid.layers.reduce_all (ArgSpec(args=['input', 'dim', 'keep_dim', 'name'], varargs=None, keywords=None, defaults=(None, False, None)), ('document', '8ab17ab51f68a6e76302b27f928cedf3'))
-paddle.fluid.layers.reduce_any (ArgSpec(args=['input', 'dim', 'keep_dim', 'name'], varargs=None, keywords=None, defaults=(None, False, None)), ('document', '0483ac3b7a99e879ccde583ae8d7a60d'))
+paddle.fluid.layers.reduce_all (ArgSpec(args=['input', 'dim', 'keep_dim', 'name'], varargs=None, keywords=None, defaults=(None, False, None)), ('document', 'e365c540ec448a73e2b2e75796ed029e'))
+paddle.fluid.layers.reduce_any (ArgSpec(args=['input', 'dim', 'keep_dim', 'name'], varargs=None, keywords=None, defaults=(None, False, None)), ('document', 'fbc9e73da7a2964ba5693864aed36abb'))
 paddle.fluid.layers.sequence_first_step (ArgSpec(args=['input'], varargs=None, keywords=None, defaults=None), ('document', '227a75392ae194de0504f5c6812dade9'))
 paddle.fluid.layers.sequence_last_step (ArgSpec(args=['input'], varargs=None, keywords=None, defaults=None), ('document', '34372f58331247749e8b0a1663cf233b'))
 paddle.fluid.layers.sequence_slice (ArgSpec(args=['input', 'offset', 'length', 'name'], varargs=None, keywords=None, defaults=(None,)), ('document', '39fbc5437be389f6c0c769f82fc1fba2'))
@@ -245,7 +245,7 @@ paddle.fluid.layers.pad2d (ArgSpec(args=['input', 'paddings', 'mode', 'pad_value
 paddle.fluid.layers.unstack (ArgSpec(args=['x', 'axis', 'num'], varargs=None, keywords=None, defaults=(0, None)), ('document', 'b0c4ca08d4eb295189e1b107c920d093'))
 paddle.fluid.layers.sequence_enumerate (ArgSpec(args=['input', 'win_size', 'pad_value', 'name'], varargs=None, keywords=None, defaults=(0, None)), ('document', 'b870fed41abd2aecf929ece65f555fa1'))
 paddle.fluid.layers.unique (ArgSpec(args=['x', 'dtype'], varargs=None, keywords=None, defaults=('int32',)), ('document', 'cab0b06e5683875f12f0efc62fa230a9'))
-paddle.fluid.layers.unique_with_counts (ArgSpec(args=['x', 'dtype'], varargs=None, keywords=None, defaults=('int32',)), ('document', '1cb59c65b41766116944b8ed1e6ad345'))
+paddle.fluid.layers.unique_with_counts (ArgSpec(args=['x', 'dtype'], varargs=None, keywords=None, defaults=('int32',)), ('document', '4496682f302007019e458a2f30d8a7c3'))
 paddle.fluid.layers.expand (ArgSpec(args=['x', 'expand_times', 'name'], varargs=None, keywords=None, defaults=(None,)), ('document', 'e93a1b102ab64b247c1b774e60d4c0d0'))
 paddle.fluid.layers.sequence_concat (ArgSpec(args=['input', 'name'], varargs=None, keywords=None, defaults=(None,)), ('document', 'f47f9d207ac60b6f294087bcb1b64ae8'))
 paddle.fluid.layers.scale (ArgSpec(args=['x', 'scale', 'bias', 'bias_after_scale', 'act', 'name'], varargs=None, keywords=None, defaults=(1.0, 0.0, True, None, None)), ('document', '463e4713806e5adaa4d20a41e2218453'))
@@ -266,7 +266,7 @@ paddle.fluid.layers.sum (ArgSpec(args=['x'], varargs=None, keywords=None, defaul
 paddle.fluid.layers.slice (ArgSpec(args=['input', 'axes', 'starts', 'ends'], varargs=None, keywords=None, defaults=None), ('document', '8c622791994a0d657d8c6c9cefa5bf34'))
 paddle.fluid.layers.strided_slice (ArgSpec(args=['input', 'axes', 'starts', 'ends', 'strides'], varargs=None, keywords=None, defaults=None), ('document', '340d8d656272ea396b441aab848429a2'))
 paddle.fluid.layers.shape (ArgSpec(args=['input'], varargs=None, keywords=None, defaults=None), ('document', 'bf61c8f79d795a8371bdb3b5468aa82b'))
-paddle.fluid.layers.rank (ArgSpec(args=['input'], varargs=None, keywords=None, defaults=None), ('document', '096df0e0273145ab80ed119a4c294db3'))
+paddle.fluid.layers.rank (ArgSpec(args=['input'], varargs=None, keywords=None, defaults=None), ('document', 'a4492cf0393c6f70e4e25c681dcd73f4'))
 paddle.fluid.layers.size (ArgSpec(args=['input'], varargs=None, keywords=None, defaults=None), ('document', 'cf2e156beae36378722666c4c33bebfe'))
 paddle.fluid.layers.logical_and (ArgSpec(args=['x', 'y', 'out', 'name'], varargs=None, keywords=None, defaults=(None, None)), ('document', '12db97c6c459c0f240ec7006737174f2'))
 paddle.fluid.layers.logical_or (ArgSpec(args=['x', 'y', 'out', 'name'], varargs=None, keywords=None, defaults=(None, None)), ('document', '15adbc561618b7db69671e02009bea67'))
@@ -303,8 +303,8 @@ paddle.fluid.layers.npair_loss (ArgSpec(args=['anchor', 'positive', 'labels', 'l
 paddle.fluid.layers.pixel_shuffle (ArgSpec(args=['x', 'upscale_factor'], varargs=None, keywords=None, defaults=None), ('document', '7e5cac851fd9bad344230e1044b6a565'))
 paddle.fluid.layers.fsp_matrix (ArgSpec(args=['x', 'y'], varargs=None, keywords=None, defaults=None), ('document', '3a4eb7cce366f5fd8bc38b42b6af5ba1'))
 paddle.fluid.layers.continuous_value_model (ArgSpec(args=['input', 'cvm', 'use_cvm'], varargs=None, keywords=None, defaults=(True,)), ('document', 'b335b531931cc8b2d19c65980eadfc1e'))
-paddle.fluid.layers.where (ArgSpec(args=['condition'], varargs=None, keywords=None, defaults=None), ('document', 'b1e1487760295e1ff55307b880a99e18'))
-paddle.fluid.layers.sign (ArgSpec(args=['x'], varargs=None, keywords=None, defaults=None), ('document', 'b56afe9ae3fc553c95d907fd7ef6c314'))
+paddle.fluid.layers.where (ArgSpec(args=['condition'], varargs=None, keywords=None, defaults=None), ('document', '68810eedf448f2cb3abd46518dd46c39'))
+paddle.fluid.layers.sign (ArgSpec(args=['x'], varargs=None, keywords=None, defaults=None), ('document', '9f19288d9a8dabcfd0bbb4fc032fa521'))
 paddle.fluid.layers.deformable_conv (ArgSpec(args=['input', 'offset', 'mask', 'num_filters', 'filter_size', 'stride', 'padding', 'dilation', 'groups', 'deformable_groups', 'im2col_step', 'param_attr', 'bias_attr', 'modulated', 'name'], varargs=None, keywords=None, defaults=(1, 0, 1, None, None, None, None, None, True, None)), ('document', '3e090f9e90b9c24d07348243bf137b56'))
 paddle.fluid.layers.unfold (ArgSpec(args=['x', 'kernel_sizes', 'strides', 'paddings', 'dilations', 'name'], varargs=None, keywords=None, defaults=(1, 0, 1, None)), ('document', '3f884662ad443d9ecc2b3734b4f61ad6'))
 paddle.fluid.layers.deformable_roi_pooling (ArgSpec(args=['input', 'rois', 'trans', 'no_trans', 'spatial_scale', 'group_size', 'pooled_height', 'pooled_width', 'part_size', 'sample_per_part', 'trans_std', 'position_sensitive', 'name'], varargs=None, keywords=None, defaults=(False, 1.0, [1, 1], 1, 1, None, 1, 0.1, False, None)), ('document', 'e0e7bf35da2287efb015546f1b8350df'))
@@ -339,10 +339,10 @@ paddle.fluid.layers.has_inf (ArgSpec(args=['x'], varargs=None, keywords=None, de
 paddle.fluid.layers.has_nan (ArgSpec(args=['x'], varargs=None, keywords=None, defaults=None), ('document', '129cf426e71452fe8276d616a6dc21ae'))
 paddle.fluid.layers.isfinite (ArgSpec(args=['x'], varargs=None, keywords=None, defaults=None), ('document', 'b9fff4ffc8d11934cde099f4c39bf841'))
 paddle.fluid.layers.range (ArgSpec(args=['start', 'end', 'step', 'dtype'], varargs=None, keywords=None, defaults=None), ('document', '3e982b788b95f959eafeeb0696a3cbde'))
-paddle.fluid.layers.linspace (ArgSpec(args=['start', 'stop', 'num', 'dtype'], varargs=None, keywords=None, defaults=None), ('document', '3663d1148946eed4c1c34c81be586b9e'))
-paddle.fluid.layers.zeros_like (ArgSpec(args=['x', 'out'], varargs=None, keywords=None, defaults=(None,)), ('document', 'd88a23bcdc443719b3953593f7cef14a'))
+paddle.fluid.layers.linspace (ArgSpec(args=['start', 'stop', 'num', 'dtype'], varargs=None, keywords=None, defaults=None), ('document', '156e653497804566a43f6a53d48b08c4'))
+paddle.fluid.layers.zeros_like (ArgSpec(args=['x', 'out'], varargs=None, keywords=None, defaults=(None,)), ('document', '5432543db3ff898451aa3af6bb38ab56'))
 paddle.fluid.layers.ones_like (ArgSpec(args=['x', 'out'], varargs=None, keywords=None, defaults=(None,)), ('document', 'd18d42059c6b189cbd3fab2fcb206c15'))
-paddle.fluid.layers.diag (ArgSpec(args=['diagonal'], varargs=None, keywords=None, defaults=None), ('document', '88a15e15f0098d549f07a01eaebf9ce3'))
+paddle.fluid.layers.diag (ArgSpec(args=['diagonal'], varargs=None, keywords=None, defaults=None), ('document', '10f13aae2f0d3db88353d3ec7db4869b'))
 paddle.fluid.layers.eye (ArgSpec(args=['num_rows', 'num_columns', 'batch_shape', 'dtype'], varargs=None, keywords=None, defaults=(None, None, 'float32')), ('document', '60cdc70ae43ba69fae36d720ef3016a1'))
 paddle.fluid.layers.While ('paddle.fluid.layers.control_flow.While', ('document', '50110155608a00f43d3d3fd1be41dcb4'))
 paddle.fluid.layers.While.__init__ (ArgSpec(args=['self', 'cond', 'is_test', 'name'], varargs=None, keywords=None, defaults=(False, None)), ('document', '6adf97f83acf6453d4a6a4b1070f3754'))
@@ -355,11 +355,11 @@ paddle.fluid.layers.increment (ArgSpec(args=['x', 'value', 'in_place'], varargs=
 paddle.fluid.layers.array_write (ArgSpec(args=['x', 'i', 'array'], varargs=None, keywords=None, defaults=(None,)), ('document', '3f913b5069ad40bd85d89b33e4aa5939'))
 paddle.fluid.layers.create_array (ArgSpec(args=['dtype'], varargs=None, keywords=None, defaults=None), ('document', '556de793fdf24d515f3fc91260e2c048'))
 paddle.fluid.layers.less_than (ArgSpec(args=['x', 'y', 'force_cpu', 'cond'], varargs=None, keywords=None, defaults=(None, None)), ('document', '04af32422c3a3d8f6040aeb406c82768'))
-paddle.fluid.layers.less_equal (ArgSpec(args=['x', 'y', 'cond'], varargs=None, keywords=None, defaults=(None,)), ('document', '7b6d952a9f6340a044cfb91c16aad842'))
-paddle.fluid.layers.greater_than (ArgSpec(args=['x', 'y', 'cond'], varargs=None, keywords=None, defaults=(None,)), ('document', '55710e2fafeda70cd1b53d7509712499'))
-paddle.fluid.layers.greater_equal (ArgSpec(args=['x', 'y', 'cond'], varargs=None, keywords=None, defaults=(None,)), ('document', '14bff27b2be5e60eaa30e41925265beb'))
+paddle.fluid.layers.less_equal (ArgSpec(args=['x', 'y', 'cond'], varargs=None, keywords=None, defaults=(None,)), ('document', '04e5623dd39b4437b9b08e0ce11071ca'))
+paddle.fluid.layers.greater_than (ArgSpec(args=['x', 'y', 'cond'], varargs=None, keywords=None, defaults=(None,)), ('document', '135352e24251238122bb7823dd4a49aa'))
+paddle.fluid.layers.greater_equal (ArgSpec(args=['x', 'y', 'cond'], varargs=None, keywords=None, defaults=(None,)), ('document', '44bdacd11299d72c0a52d2181e7ae6ca'))
 paddle.fluid.layers.equal (ArgSpec(args=['x', 'y', 'cond'], varargs=None, keywords=None, defaults=(None,)), ('document', '788aa651e8b9fec79d16931ef3a33e90'))
-paddle.fluid.layers.not_equal (ArgSpec(args=['x', 'y', 'cond'], varargs=None, keywords=None, defaults=(None,)), ('document', '57adebb8858ffab6be2d86d0522b85dc'))
+paddle.fluid.layers.not_equal (ArgSpec(args=['x', 'y', 'cond'], varargs=None, keywords=None, defaults=(None,)), ('document', '8b76aaac4ba7cf9111750b9c2c9418cb'))
 paddle.fluid.layers.array_read (ArgSpec(args=['array', 'i'], varargs=None, keywords=None, defaults=None), ('document', 'caf0d94349cdc28e1bda3b8a19411ac0'))
 paddle.fluid.layers.array_length (ArgSpec(args=['array'], varargs=None, keywords=None, defaults=None), ('document', '6f24a9b872027634ad758ea2826c9727'))
 paddle.fluid.layers.IfElse ('paddle.fluid.layers.control_flow.IfElse', ('document', 'a389f88e19c3b332c3afcbf7df4488a5'))

python/paddle/fluid/layers/control_flow.py

@@ -1042,24 +1042,28 @@ def less_than(x, y, force_cpu=None, cond=None):
 @templatedoc()
 def less_equal(x, y, cond=None):
     """
-    This layer returns the truth value of :math:`x <= y` elementwise, which is equivalent to the overloaded operator `<=`.
+    This OP returns the truth value of :math:`x <= y` elementwise, which is equivalent function to the overloaded operator `<=`.
 
     Args:
-        x(Variable): First operand of *less_equal*
-        y(Variable): Second operand of *less_equal*
-        cond(Variable|None): Optional output variable to store the result of *less_equal*
+        x(Variable): First input to compare which is N-D tensor. The input data type should be float32, float64, int32, int64. 
+        y(Variable): Second input to compare which is N-D tensor. The input data type should be float32, float64, int32, int64.
+        cond(Variable, optional): If is :attr:`None`, the op will create a variable as output tensor, the input shape and data type of \
+            this tensor is the same as input :attr:`x`. If is not :attr:`None`, the op will set the variable as output tensor, the input shape \
+            and data type of this tensor should be the same as input :attr:`x`. Default value is :attr:`None`.
 
     Returns:
-        Variable: The tensor variable storing the output of *less_equal*.
+        Variable, the output data type is bool.: The tensor variable storing the output, the output shape is the same as input :attr:`x`.
 
     Examples:
         .. code-block:: python
 
           import paddle.fluid as fluid
+          import numpy as np
+          label = fluid.layers.assign(np.array([1, 3], dtype='int32'))
+          limit = fluid.layers.assign(np.array([1, 2], dtype='int32'))
+          out = fluid.layers.less_equal(x=label, y=limit) #out=[True, False]
+          out1 = label<= limit #out1=[True, False]
 
-          label = fluid.layers.data(name='label', shape=[1], dtype='int64')
-          limit = fluid.layers.fill_constant(shape=[1], value=1, dtype='int64')
-          out = fluid.layers.less_equal(x=label, y=limit)
     """
     helper = LayerHelper("less_equal", **locals())
     if cond is None:
@@ -1082,24 +1086,27 @@ def less_equal(x, y, cond=None):
 @templatedoc()
 def greater_than(x, y, cond=None):
     """
-    This layer returns the truth value of :math:`x > y` elementwise, which is equivalent to the overloaded operator `>`.
+    This OP returns the truth value of :math:`x > y` elementwise, which is equivalent function to the overloaded operator `>`.
 
     Args:
-        x(Variable): First operand of *greater_than*
-        y(Variable): Second operand of *greater_than*
-        cond(Variable|None): Optional output variable to store the result of *greater_than*
+        x(Variable): First input to compare which is N-D tensor. The input data type should be float32, float64, int32, int64. 
+        y(Variable): Second input to compare which is N-D tensor. The input data type should be float32, float64, int32, int64.
+        cond(Variable, optional): If is :attr:`None`, the op will create a variable as output tensor, the shape and data type of this \
+            tensor is the same as input :attr:`x` . If is not :attr:`None`, the op will set the variable as output tensor, the shape and data type \
+            of this tensor should be the same as input :attr:`x` . Default value is :attr:`None`.
 
     Returns:
-        Variable: The tensor variable storing the output of *greater_than*.
+        Variable, the output data type is bool.: The tensor variable storing the output, the output shape is the same as input :attr:`x` .
 
     Examples:
         .. code-block:: python
 
           import paddle.fluid as fluid
-          
-          label = fluid.layers.data(name='label', shape=[1], dtype='int64')
-          limit = fluid.layers.fill_constant(shape=[1], value=1, dtype='int64')
-          out = fluid.layers.greater_than(x=label, y=limit)
+          import numpy as np
+          label = fluid.layers.assign(np.array([2, 3], dtype='int32'))
+          limit = fluid.layers.assign(np.array([3, 2], dtype='int32'))
+          out = fluid.layers.greater_than(x=label, y=limit) #out=[False, True]
+          out1 = label > limit #out1=[False, True]
     """
     helper = LayerHelper("greater_than", **locals())
     if cond is None:
@@ -1122,24 +1129,28 @@ def greater_than(x, y, cond=None):
 @templatedoc()
 def greater_equal(x, y, cond=None):
     """
-    This layer returns the truth value of :math:`x >= y` elementwise, which is equivalent to the overloaded operator `>=`.
+    This OP returns the truth value of :math:`x >= y` elementwise, which is equivalent function to the overloaded operator `>=`.
 
     Args:
-        x(Variable): First operand of *greater_equal*
-        y(Variable): Second operand of *greater_equal*
-        cond(Variable|None): Optional output variable to store the result of *greater_equal*
+        x(Variable): First input to compare which is N-D tensor. The input data type should be float32, float64, int32, int64. 
+        y(Variable): Second input to compare which is N-D tensor. The input data type should be float32, float64, int32, int64.
+        cond(Variable, optional): If is :attr:`None` , the op will create a variable as output tensor, the shape and data type of this \
+            tensor is the same as input :attr:`x`. If is not :attr:`None` , the op will set the variable as output tensor, the shape and data \
+            type of this tensor is the same as input :attr:`x`. Default value is :attr:`None`.
 
     Returns:
-        Variable: The tensor variable storing the output of *greater_equal*.
+        Variable, the output data type is bool.: The tensor variable storing the output, the output shape is the same as input :attr:`x`.
 
     Examples:
         .. code-block:: python
 
           import paddle.fluid as fluid
+          import numpy as np
 
-          label = fluid.layers.data(name='label', shape=[1], dtype='int64')
-          limit = fluid.layers.fill_constant(shape=[1], value=1, dtype='int64')
-          out = fluid.layers.greater_equal(x=label, y=limit)
+          label = fluid.layers.assign(np.array([2, 2], dtype='int32'))
+          limit = fluid.layers.assign(np.array([2, 3], dtype='int32'))
+          out = fluid.layers.greater_equal(x=label, y=limit) #out=[True, False]
+          out_1 = label >= limit #out1=[True, False]
 
     """
     helper = LayerHelper("greater_equal", **locals())
@@ -1193,15 +1204,17 @@ def equal(x, y, cond=None):
 
 def not_equal(x, y, cond=None):
     """
-    This layer returns the truth value of :math:`x != y` elementwise, which is equivalent to the overloader operator `!=`.
+    This OP returns the truth value of :math:`x != y` elementwise, which is equivalent function to the overloaded operator `!=`.
 
     Args:
-        x(Variable): First operand of *not_equal*
-        y(Variable): Second operand of *not_equal*
-        cond(Variable|None): Optional output variable to store the result of *not_equal*
+        x(Variable): First input to compare which is N-D tensor. The input data type should be float32, float64, int32, int64. 
+        y(Variable): Second input to compare which is N-D tensor. The input data type should be float32, float64, int32, int64.
+        cond(Variable, optional): If is :attr:`None`, the op will create a variable as output tensor, the shape and data type of this \
+             tensor is the same as input :attr:`x`. If is not :attr:`None`, the op will set the variable as output tensor, the shape and data \
+             type of this tensor should be the same as input :attr:`x`. Default value is :attr:`None`.
 
     Returns:
-        Variable: The tensor variable storing the output of *not_equal*.
+        Variable, the output data type is bool.: The tensor variable storing the output, the output shape is the same as input :attr:`x`.
 
     Examples:
         .. code-block:: python

python/paddle/fluid/layers/nn.py

@@ -6092,23 +6092,23 @@ def reduce_prod(input, dim=None, keep_dim=False, name=None):
 
 def reduce_all(input, dim=None, keep_dim=False, name=None):
     """
-    Computes the ``logical and`` of tensor elements over the given dimension.
+    This OP computes the ``logical and`` of tensor elements over the given dimension, and output the result.
 
     Args:
-        input (Variable): The input variable which is a Tensor or LoDTensor.
-        dim (list|int|None): The dimension along which the logical and is computed.
+        input (Variable): The input variable which is a Tensor or LoDTensor, the input data type should be `bool`.
+        dim (list|int|optional): The dimension along which the logical and is computed.
             If :attr:`None`, compute the logical and over all elements of
             :attr:`input` and return a Tensor variable with a single element,
             otherwise must be in the range :math:`[-rank(input), rank(input))`.
-            If :math:`dim[i] < 0`, the dimension to reduce is :math:`rank + dim[i]`.
+            If :math:`dim[i] < 0`, the dimension to reduce is :math:`rank + dim[i]`. The default value is None. 
         keep_dim (bool): Whether to reserve the reduced dimension in the
             output Tensor. The result tensor will have one fewer dimension
-            than the :attr:`input` unless :attr:`keep_dim` is true.
+            than the :attr:`input` unless :attr:`keep_dim` is true. The default value is False.
         name(str|None): A name for this layer(optional). If set None, the layer
-                       will be named automatically.
+                       will be named automatically. The default value is None. 
 
     Returns: 
-        Variable: The reduced Tensor variable.
+        Variable, the output data type is bool. : The reduced tensor variable with ``logical and`` in given dims.
 
     Examples:
         .. code-block:: python
@@ -6126,7 +6126,10 @@ def reduce_all(input, dim=None, keep_dim=False, name=None):
             out = layers.reduce_all(x)  # False 
             out = layers.reduce_all(x, dim=0)  # [True, False]
             out = layers.reduce_all(x, dim=-1)  # [False, True]
+            # keep_dim=False, x.shape=(2,2), out.shape=(2,)
+
             out = layers.reduce_all(x, dim=1, keep_dim=True)  # [[False], [True]]
+            # keep_dim=True, x.shape=(2,2), out.shape=(2,1)
 
     """
     helper = LayerHelper('reduce_all', **locals())
@@ -6147,23 +6150,22 @@ def reduce_all(input, dim=None, keep_dim=False, name=None):
 
 def reduce_any(input, dim=None, keep_dim=False, name=None):
     """
-    Computes the ``logical or`` of tensor elements over the given dimension.
+    This OP computes the ``logical or`` of tensor elements over the given dimension, and output the result.
 
     Args:
-        input (Variable): The input variable which is a Tensor or LoDTensor.
-        dim (list|int|None): The dimension along which the logical or is computed.
-            If :attr:`None`, compute the logical or over all elements of
+        input (Variable): The input variable which is a Tensor or LoDTensor, the input data type should be `bool`.
+        dim (list|int|optional): The dimension along which the logical and is computed.
+            If :attr:`None`, compute the logical and over all elements of
             :attr:`input` and return a Tensor variable with a single element,
             otherwise must be in the range :math:`[-rank(input), rank(input))`.
-            If :math:`dim[i] < 0`, the dimension to reduce is :math:`rank + dim[i]`.
+            If :math:`dim[i] < 0`, the dimension to reduce is :math:`rank + dim[i]`. The default value is None. 
         keep_dim (bool): Whether to reserve the reduced dimension in the
             output Tensor. The result tensor will have one fewer dimension
-            than the :attr:`input` unless :attr:`keep_dim` is true.
+            than the :attr:`input` unless :attr:`keep_dim` is true. The default value is False.
         name(str|None): A name for this layer(optional). If set None, the layer
-                       will be named automatically.
 
     Returns: 
-        Variable: The reduced Tensor variable.
+        Variable, the output data type is bool. : The reduced tensor variable with ``logical or`` in given dims.
 
     Examples:
         .. code-block:: python
@@ -6181,8 +6183,11 @@ def reduce_any(input, dim=None, keep_dim=False, name=None):
             out = layers.reduce_any(x)  # True
             out = layers.reduce_any(x, dim=0)  # [True, False]
             out = layers.reduce_any(x, dim=-1)  # [True, False]
+            # keep_dim=False, x.shape=(2,2), out.shape=(2,)
+
             out = layers.reduce_any(x, dim=1,
                                      keep_dim=True)  # [[True], [False]]
+            # keep_dim=True, x.shape=(2,2), out.shape=(2,1)
 
     """
     helper = LayerHelper('reduce_any', **locals())
@@ -12535,23 +12540,21 @@ def shape(input):
 
 def rank(input):
     """
-    **Rank Layer**
-
-    Returns the number of dimensions for a tensor, which is a 0-D int32 Tensor.
+    The OP returns the number of dimensions for a tensor, which is a 0-D int32 Tensor.
 
     Args:
-        input (Variable): The input variable.
+        input (Variable): The input N-D tensor with shape of :math:`[N_1, N_2, ..., N_k]`, the data type is arbitrary.
 
     Returns:
-        Variable: The rank of the input variable.
+        Variable, the output data type is int32.: The 0-D tensor with the dimensions of the input variable.
 
     Examples:
         .. code-block:: python
 
             import paddle.fluid as fluid
 
-            input = fluid.layers.data(name="input", shape=[3, 100, 100], dtype="float32")
-            rank = fluid.layers.rank(input) # 4
+            input = fluid.data(name="input", shape=[3, 100, 100], dtype="float32")
+            rank = fluid.layers.rank(input) # rank=(3,)
     """
 
     ndims = len(input.shape)
@@ -15203,14 +15206,11 @@ def where(condition):
     """
     Return an int64 tensor with rank 2, specifying the coordinate of true element in `condition`.
 
-    Output's first dimension is the number of true element, second dimension is rank(number of dimension) of `condition`.
-    If there is zero true element, then an empty tensor will be generated.  
-
     Args:
-        condition(Variable): A bool tensor with rank at least 1.
+        condition(Variable): A bool tensor with rank at least 1, the data type is bool.
 
     Returns:
-        Variable: The tensor variable storing a 2-D tensor. 
+        Variable, the output data type is int64. : The tensor variable storing a 2-D tensor, which involves all coordinate. 
 
     Examples:
         .. code-block:: python
@@ -15247,15 +15247,14 @@ def where(condition):
 
 def sign(x):
     """
-    **sign**
-
-    This function returns sign of every element in `x`: 1 for positive, -1 for negative and 0 for zero.
+    This OP returns sign of every element in `x`: 1 for positive, -1 for negative and 0 for zero.
 
     Args:
-        x(Variable|numpy.ndarray): The input tensor.
+        x(Variable|numpy.ndarray): The input variable could be N-D tensor or N-D numpy array, \
+            the input data type is float32 or float64.
 
     Returns:
-        Variable: The output sign tensor with identical shape and dtype to `x`.
+        Variable, the output data type is the same as input data type. : The output sign tensor with identical shape to input :attr:`x`.
 
     Examples:
         .. code-block:: python
@@ -15263,9 +15262,8 @@ def sign(x):
           import paddle.fluid as fluid
           import numpy as np
 
-          # [1, 0, -1]
-          data = fluid.layers.sign(np.array([3, 0, -2], dtype='float32')) 
-
+          # [1.0, 0.0, -1.0]
+          data = fluid.layers.sign(np.array([3.0, 0.0, -2.0], dtype='float32')) 
     """
 
     helper = LayerHelper("sign", **locals())
@@ -15331,18 +15329,21 @@ def unique(x, dtype='int32'):
 
 def unique_with_counts(x, dtype='int32'):
     """
-    **unique** 
+    This OP return a unique tensor for `x` , and count tensor that the count of unqiue result in raw input, \
+    and an index tensor pointing to this unique tensor. 
 
-    Return a unique tensor for `x` and an index tensor pointing to this unique tensor.
+    **NOTICE**: This op just be supported in device of CPU, and support the variable type of Tensor only.
 
     Args:
-        x(Variable): A 1-D input tensor.
-        dtype(np.dtype|core.VarDesc.VarType|str): The type of index tensor: int32, int64.
+        x(Variable): A 1-D input tensor with input shape of :math:`[N]` , the input data type is float32, float64, int32, int64.
+        dtype(np.dtype|core.VarDesc.VarType|str): The type of count and index tensor, it could be int32, int64. Defalut value is int32.
 
     Returns: 
-        tuple: (out, index, count). `out` is the unique tensor for `x`, with identical dtype to `x`, and \
-            `index` is an index tensor pointing to `out`, by which user can recover the original `x` tensor, \
-            `count` is count of unqiue element in the `x`.
+        tuple, the variable type in tuple is Tensor, the output :attr:`out` data type is the same as input :attr:`x`, \
+        and data type of output :attr:`index` and :attr:`count` will be int32 or int64.: The :attr:`out` is unique tensor for input :attr:`x`,\
+        the data shape is :math:`[K]`, the `K` may be different to the `N` in shape of :attr:`x`. :attr:`index` is an index tensor pointing\
+        to :attr:`out`, the data shape is :math:`[N]` , the data shape is the same as input :attr:`x`. :attr:`count` is count of unqiue element in\
+        the :attr:`x`, the data shape is :math:`[K]`, the data shape is the same as output :attr:`out`.
 
     Examples:
         .. code-block:: python
@@ -15352,6 +15353,7 @@ def unique_with_counts(x, dtype='int32'):
              x = fluid.layers.assign(np.array([2, 3, 3, 1, 5, 3], dtype='int32'))
              out, index, count = fluid.layers.unique_with_counts(x) # out is [2, 3, 1, 5]; index is [0, 1, 1, 2, 3, 1]
                                                         # count is [1, 3, 1, 1]
+            # x.shape=(6,) out.shape=(4,), index.shape=(6,), count.shape=(4,)
     """
     if not (dtype == 'int32' or dtype == 'int64'):
         raise TypeError(

python/paddle/fluid/layers/tensor.py

@@ -896,18 +896,21 @@ def range(start, end, step, dtype):
 
 def linspace(start, stop, num, dtype):
     """
-    Return fixed number of evenly spaced values within a given interval.
-
-    First entry is start, and last entry is stop. In the case when Num is 1, only Start is returned. Like linspace function of numpy.
+    This OP return fixed number of evenly spaced values within a given interval.
 
     Args:
-        start(float|Variable): First entry in the sequence. It is a float scalar, or a tensor of shape [1] with type 'float32'|'float64'.
-        stop(float|Variable): Last entry in the sequence. It is a float scalar, or a tensor of shape [1] with type 'float32'|'float64'.
-        num(int|Variable): Number of entry in the sequence. It is an int scalar, or a tensor of shape [1] with type int32.
-        dtype(string): 'float32'|'float64', the data type of the output tensor.
+        start(float|Variable): The input :attr:`start` is start variable of range. It is a float scalar, \
+            or a tensor of shape [1] with input data type float32, float64.
+        stop(float|Variable): The input :attr:`stop` is start variable of range. It is a float scalar, \
+            or a tensor of shape [1] with input data type float32, float64.
+        num(int|Variable): The input :attr:`num` is given num of the sequence. It is an int scalar, \
+            or a tensor of shape [1] with type int32.
+        dtype(string): The data type of output tensor, it could be 'float32' and 'float64'.
 
     Returns:
-        Variable: The tensor variable storing a 1-D tensor. 
+        Variable, the output data type will be float32, float64.: The 1-D tensor with fixed number of evenly spaced values, \
+        the data shape of this tensor is :math:`[num]` . If the :attr:`num` is set 1, the output tensor just has \
+        the value with input :attr:`start`. 
 
     Examples:
         .. code-block:: python
@@ -939,23 +942,24 @@ def linspace(start, stop, num, dtype):
 
 def zeros_like(x, out=None):
     """
-    **zeros_like**
-
-    This function creates a zeros tensor which has identical shape and dtype 
+    This OP creates a zeros tensor which has identical shape and dtype 
     with `x`.
 
     Args:
-        x(Variable): The input tensor which specifies shape and dtype.
-        out(Variable): The output tensor.
+        x(Variable): The input tensor which specifies shape and dtype, the input data dtype could be bool, float32, float64, int32, int64.
+        out(Variable, optional): If is :attr:`None` , the op will create the variable as output, the data type and shape of \
+            this variable will be same as input :attr:`x`. If is a tensor, the data type and shape need to be same as input :attr:`x`. 
+            The defalut value is :attr:`None` .
 
     Returns:
-        Variable: The tensor variable storing the output.
+        Variable: The N-D tensor, the element in tensor is related to input data type, if the input data type is bool, \
+            the output value is False, otherwise is zero. The output shape is the same as the input.
 
     Examples:
         .. code-block:: python
 
           import paddle.fluid as fluid
-          x = fluid.layers.data(name='x', dtype='float32', shape=[3], append_batch_size=False)
+          x = fluid.data(name='x', dtype='float32', shape=[3])
           data = fluid.layers.zeros_like(x) # [0.0, 0.0, 0.0]
 
     """
@@ -971,15 +975,15 @@ def zeros_like(x, out=None):
 
 def diag(diagonal):
     """
-    **diag**
-
-    This function creates a square matrix which has diagonal values specified by `diagonal`.
+    This OP creates a square matrix which has diagonal values specified by input :attr:`diagonal`.
 
     Args:
-        diagonal(Variable|numpy.ndarray): The input tensor specifying diagonal values, should be of rank 1.
+        diagonal(Variable|numpy.ndarray): The input tensor should be 1D tensor, the input shape is :math:`[ N]` , \
+            specifying diagonal values by this input tensor. The input data type should be float32, float64, int32, int64.
 
     Returns:
-        Variable: The tensor variable storing the square matrix.
+        Variable, the output data type is the same as input data type.: The tensor variable storing the square matrix, \
+            the diagonal values specified by input :attr:`diagonal`. the output shape is :math:`[N, N]` with two dims.
 
     Examples:
         .. code-block:: python
@@ -990,7 +994,9 @@ def diag(diagonal):
 
           import paddle.fluid as fluid
           import numpy as np
-          data = fluid.layers.diag(np.arange(3, 6, dtype='int32')) 
+          diagonal = np.arange(3, 6, dtype='int32')
+          data = fluid.layers.diag(diagonal)
+          # diagonal.shape=(3,) data.shape=(3, 3)
 
     """