modify code display way

docs/api/api_python/mindspore.rewrite.rst

@@ -6,6 +6,10 @@ ReWrite完整示例请参考
 `rewrite_example.py <https://gitee.com/mindspore/mindspore/tree/master/docs/api/api_python/rewrite_example.py>`_ 。
 该样例代码的主要功能包括：怎么通过网络创建SymbolTree，并且对SymbolTree中的节点进行插入删除替换等操作，其中还包含了对子网络的修改和通过模式匹配进行节点替换。
 
+.. literalinclude:: rewrite_example.py
+    :language: python
+    :start-at: import
+
 .. py:class:: mindspore.rewrite.SymbolTree(handler: SymbolTreeImpl)
 
     SymbolTree通常对应于网络的前向计算过程。
@@ -483,7 +487,7 @@ ReWrite完整示例请参考
             - **RuntimeError** - 如果参数 `node` 不是 NodeType.Tree类型。
             - **TypeError** - 如果参数 `node` 不是Node类型实例。
 
-.. py:method:: mindspore.rewrite.sparsify(f, arg_types, sparse_rules=None)
+.. py:function:: mindspore.rewrite.sparsify(f, arg_types, sparse_rules=None)
 
     模型自动稀疏化接口，将稠密模型转换为稀疏模型。通过 `arg_types` 指定的参数类型，将稀疏参数在模型中传导，并调用相应的稀疏函数。
 

docs/api/api_python/nn/mindspore.nn.thor.rst

@@ -47,7 +47,7 @@ mindspore.nn.thor
         - **decay_filter** (function) - 用于确定权重衰减应用于哪些层的函数，只有在weight_decay>0时才有效。默认值：lambda x: x.name not in []。
         - **split_indices** (list) - 按A/G层（A/G含义见上述公式）索引设置allreduce融合策略。仅在分布式计算中有效。ResNet50作为一个样本，A/G的层数分别为54层，当split_indices设置为[26,53]时，表示A/G被分成两组allreduce，一组为0~26层，另一组是27~53层。默认值：None。
         - **enable_clip_grad** (bool) - 是否剪切梯度。默认值：False。
-        - **frequency** (int) - A/G和$A^{-1}/G^{-1}$的更新间隔。每隔frequency个step，A/G和$A^{-1}/G^{-1}$将更新一次。必须大于1。默认值：100。
+        - **frequency** (int) - A/G和 :math:`A^{-1}/G^{-1}` 的更新间隔。当frequency等于N(N必须大于1)，每隔frequency个step，A/G和 :math:`A^{-1}/G^{-1}` 将更新一次。其他step将使用之前的A/G和 :math:`A^{-1}/G^{-1}` 来更新权重。默认值：100。
 
     输入：
         - **gradients** （tuple[Tensor]） - 训练参数的梯度，矩阵维度与训练参数相同。

docs/api/api_python_en/mindspore.rewrite.rst

@@ -5,5 +5,9 @@ For a complete ReWrite example, refer to
 `rewrite_example.py <https://gitee.com/mindspore/mindspore/tree/master/docs/api/api_python_en/rewrite_example.py>`_ 。
 The main functions of the sample code include: how to create a SymbolTree through the network, and how to insert, delete, and replace the nodes in the SymbolTree. It also includes the modification of the subnet and node replacement through pattern matching.
 
+.. literalinclude:: rewrite_example.py
+    :language: python
+    :start-at: import
+
 .. automodule:: mindspore.rewrite
     :members:

mindspore/ccsrc/minddata/dataset/include/dataset/data_helper.h

@@ -340,7 +340,7 @@ class DATASET_API DataHelper {
   }
 
   /// \brief Write pointer to bin, use pointer to avoid memcpy
-  /// \note The value of `length`` must be equal to the length of `data`
+  /// \note The value of `length` must be equal to the length of `data`
   /// \param[in] in_file File name to write to
   /// \param[in] data Pointer to data
   /// \param[in] length Length of values to write from pointer

mindspore/core/ops/expand_dims.h

@@ -24,7 +24,7 @@
 namespace mindspore {
 namespace ops {
 constexpr auto kNameExpandDims = "ExpandDims";
-/// \brief Adds an additional dimension to ‘input_x` at the given axis.
+/// \brief Adds an additional dimension to `input_x` at the given axis.
 /// Refer to Python API @ref mindspore.ops.ExpandDims for more details.
 class MIND_API ExpandDims : public BaseOperator {
  public:

mindspore/python/mindspore/nn/optim/thor.py

@@ -314,9 +314,9 @@ def thor(net, learning_rate, damping, momentum, weight_decay=0.0, loss_scale=1.0
 
         enable_clip_grad (bool): Whether to clip the gradients. Default: False
 
-        frequency(int): The update interval of A/G and $A^{-1}/G^{-1}$. When frequency equals N (N is greater than 1),
+        frequency(int): The update interval of A/G and :math:`A^{-1}/G^{-1}`. When frequency equals N
-            A/G and $A^{-1}/G^{-1}$ will be updated  every N steps, and other steps will use the stale A/G and
+            (N is greater than 1), A/G and :math:`A^{-1}/G^{-1}` will be updated every N steps,
-            $A^{-1}/G^{-1}$ to update weights. Default: 100.
+            and other steps will use the stale A/G and :math:`A^{-1}/G^{-1}` to update weights. Default: 100.
 
     Inputs:
         - **gradients** (tuple[Tensor]) - The gradients of `params`, the shape is the same as `params`.