move gserver codes to legacy.

This is part of big move of v2 codes to legacy sub dir.

doc/v2/dev/new_layer_cn.rst

@@ -58,7 +58,7 @@ PaddlePaddle的base layer类可以自动计算上面的导数。
 实现C++类
 ===================
 
-一个网络层的C++类需要实现初始化，前向和后向。全连接层的实现位于:code:`paddle/gserver/layers/FullyConnectedLayer.h`及:code:`paddle/gserver/layers/FullyConnectedLayer.cpp`。这里我们展示一份简化过的代码。
+一个网络层的C++类需要实现初始化，前向和后向。全连接层的实现位于:code:`paddle/legacy/gserver/layers/FullyConnectedLayer.h`及:code:`paddle/legacy/gserver/layers/FullyConnectedLayer.cpp`。这里我们展示一份简化过的代码。
 
 这个类需要继承 :code:`paddle::Layer` 这个基类，并且需要重写基类中的以下几个虚函数：
 
@@ -262,7 +262,7 @@ PaddlePaddle的base layer类可以自动计算上面的导数。
     REGISTER_LAYER(fc, FullyConnectedLayer);
     }
 
-若 :code:`cpp` 被放在 :code:`paddle/gserver/layers` 目录下，其会自动被加入编译列表。
+若 :code:`cpp` 被放在 :code:`paddle/legacy/gserver/layers` 目录下，其会自动被加入编译列表。
 
 
 写梯度检查单元测试
@@ -270,7 +270,7 @@ PaddlePaddle的base layer类可以自动计算上面的导数。
 
 写梯度检查单元测试是一个验证新实现的层是否正确的相对简单的办法。梯度检查单元测试通过有限差分法来验证一个层的梯度。首先对输入做一个小的扰动 :math:`\Delta x` ，然后观察到输出的变化为 :math:`\Delta y` ，那么，梯度就可以通过这个方程计算得到 :math:`\frac{\Delta y}{\Delta x }` 。之后，再用这个梯度去和 :code:`backward` 函数得到的梯度去对比，以保证梯度计算的正确性。需要注意的是梯度检查仅仅验证了梯度的计算，并不保证 :code:`forward` 和 :code:`backward` 函数的实现是正确的。你需要一些更复杂的单元测试来保证你实现的网络层是正确的。
 
-所有网络层的梯度检查单测都位于 :code:`paddle/gserver/tests/test_LayerGrad.cpp` 。我们建议你在写新网络层时把测试代码放入新的文件中。下面列出了全连接层的梯度检查单元测试。它包含以下几步：
+所有网络层的梯度检查单测都位于 :code:`paddle/legacy/gserver/tests/test_LayerGrad.cpp` 。我们建议你在写新网络层时把测试代码放入新的文件中。下面列出了全连接层的梯度检查单元测试。它包含以下几步：
 
 + 生成网络层配置。网络层配置包含以下几项：
    - 偏置参数的大小。（例子中是4096）
@@ -322,7 +322,7 @@ PaddlePaddle的base layer类可以自动计算上面的导数。
       }
     }
 
-如果你要为了测试而增加新的文件，例如 :code:`paddle/gserver/tests/testFCGrad.cpp` ，你需要把该文件加入 :code:`paddle/gserver/tests/CMakeLists.txt` 中。下面给出了一个例子。当你执行命令 :code:`make tests` 时，所有的单测都会被执行一次。注意，有些层可能需要高精度来保证梯度检查单测正确执行。你需要在配置cmake时将 :code:`WITH_DOUBLE` 设置为 `ON` 。
+如果你要为了测试而增加新的文件，例如 :code:`paddle/legacy/gserver/tests/testFCGrad.cpp` ，你需要把该文件加入 :code:`paddle/legacy/gserver/tests/CMakeLists.txt` 中。下面给出了一个例子。当你执行命令 :code:`make tests` 时，所有的单测都会被执行一次。注意，有些层可能需要高精度来保证梯度检查单测正确执行。你需要在配置cmake时将 :code:`WITH_DOUBLE` 设置为 `ON` 。
 
 .. code-block:: bash
 

doc/v2/dev/new_layer_en.rst

@@ -58,7 +58,7 @@ Finally we can use chain rule to calculate :math:`\frac{\partial z}{\partial x}`
 Implement C++ Class
 ===================
 
-The C++ class of the layer implements the initialization, forward, and backward part of the layer. The fully connected layer is at :code:`paddle/gserver/layers/FullyConnectedLayer.h` and :code:`paddle/gserver/layers/FullyConnectedLayer.cpp`. We list simplified version of the code below.
+The C++ class of the layer implements the initialization, forward, and backward part of the layer. The fully connected layer is at :code:`paddle/legacy/gserver/layers/FullyConnectedLayer.h` and :code:`paddle/legacy/gserver/layers/FullyConnectedLayer.cpp`. We list simplified version of the code below.
 
 It needs to derive the base class :code:`paddle::Layer`, and it needs to override the following functions:
 
@@ -263,7 +263,7 @@ Finally, you can use :code:`REGISTER_LAYER(fc, FullyConnectedLayer);` to registe
     REGISTER_LAYER(fc, FullyConnectedLayer);
     }
 
-If the :code:`cpp` file is put into :code:`paddle/gserver/layers`, it will be automatically added to the compilation list.
+If the :code:`cpp` file is put into :code:`paddle/legacy/gserver/layers`, it will be automatically added to the compilation list.
 
 
 Write Gradient Check Unit Test
@@ -271,7 +271,7 @@ Write Gradient Check Unit Test
 
 An easy way to verify the correctness of new layer's implementation is to write a gradient check unit test. Gradient check unit test utilizes finite difference method to verify the gradient of a layer. It modifies the input with a small perturbation :math:`\Delta x` and observes the changes of output :math:`\Delta y`, the gradient can be computed as :math:`\frac{\Delta y}{\Delta x }`. This gradient can be compared with the gradient computed by the :code:`backward` function of the layer to ensure the correctness of the gradient computation. Notice that the gradient check only tests the correctness of the gradient computation, it does not necessarily guarantee the correctness of the implementation of the :code:`forward` and :code:`backward` function. You need to write more sophisticated unit tests to make sure your layer is implemented correctly.
 
-All the gradient check unit tests are located in :code:`paddle/gserver/tests/test_LayerGrad.cpp`. You are recommended to put your test into a new test file if you are planning to write a new layer. The gradient test of the gradient check unit test of the fully connected layer is listed below. It has the following steps.
+All the gradient check unit tests are located in :code:`paddle/legacy/gserver/tests/test_LayerGrad.cpp`. You are recommended to put your test into a new test file if you are planning to write a new layer. The gradient test of the gradient check unit test of the fully connected layer is listed below. It has the following steps.
 
 + Create layer configuration. A layer configuration can include the following attributes:
    - size of the bias parameter. (4096 in our example)
@@ -323,7 +323,7 @@ All the gradient check unit tests are located in :code:`paddle/gserver/tests/tes
       }
     }
 
-If you are creating a new file for the test, such as :code:`paddle/gserver/tests/testFCGrad.cpp`, you need to add the file to :code:`paddle/gserver/tests/CMakeLists.txt`. An example is given below. All the unit tests will run when you execute the command :code:`make tests`. Notice that some layers might need high accuracy for the gradient check unit tests to work well. You need to configure :code:`WITH_DOUBLE` to `ON` when configuring cmake.
+If you are creating a new file for the test, such as :code:`paddle/legacy/gserver/tests/testFCGrad.cpp`, you need to add the file to :code:`paddle/legacy/gserver/tests/CMakeLists.txt`. An example is given below. All the unit tests will run when you execute the command :code:`make tests`. Notice that some layers might need high accuracy for the gradient check unit tests to work well. You need to configure :code:`WITH_DOUBLE` to `ON` when configuring cmake.
 
 .. code-block:: bash
 

doc/v2/faq/parameter/index_cn.rst

@@ -196,6 +196,6 @@ PaddlePaddle保存的模型参数文件内容由16字节头信息和网络参数
         obj="process",
         args={"src_dict_path": src_dict_path})
 
-完整源码可参考 `sequence_recurrent <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/gserver/tests/sequence_recurrent.py>`_ 示例。
+完整源码可参考 `sequence_recurrent <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/legacy/gserver/tests/sequence_recurrent.py>`_ 示例。
 
 

doc/v2/howto/rnn/hrnn_rnn_api_compare_cn.rst

@@ -4,7 +4,7 @@
 单双层RNN API对比介绍
 #####################
 
-本文以PaddlePaddle的双层RNN单元测试为示例，用多对效果完全相同的、分别使用单双层RNN作为网络配置的模型，来讲解如何使用双层RNN。本文中所有的例子，都只是介绍双层RNN的API接口，并不是使用双层RNN解决实际的问题。如果想要了解双层RNN在具体问题中的使用，请参考\ :ref:`algo_hrnn_demo`\ 。本文中示例所使用的单元测试文件是\ `test_RecurrentGradientMachine.cpp <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/gserver/tests/test_RecurrentGradientMachine.cpp>`_\ 。
+本文以PaddlePaddle的双层RNN单元测试为示例，用多对效果完全相同的、分别使用单双层RNN作为网络配置的模型，来讲解如何使用双层RNN。本文中所有的例子，都只是介绍双层RNN的API接口，并不是使用双层RNN解决实际的问题。如果想要了解双层RNN在具体问题中的使用，请参考\ :ref:`algo_hrnn_demo`\ 。本文中示例所使用的单元测试文件是\ `test_RecurrentGradientMachine.cpp <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/legacy/gserver/tests/test_RecurrentGradientMachine.cpp>`_\ 。
 
 示例1：双层RNN，子序列间无Memory
 ================================
@@ -13,8 +13,8 @@
 
 在本示例中，单层RNN和双层RNN的网络配置，都是将每一句分好词后的句子，使用LSTM作为encoder，压缩成一个向量。区别是RNN使用两层序列模型，将多句话看成一个整体同时使用encoder压缩。二者语意上完全一致。这组语义相同的示例配置如下：
 
-* 单层RNN\: `sequence_layer_group.conf <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/gserver/tests/sequence_layer_group.conf>`_
-* 双层RNN\: `sequence_nest_layer_group.conf <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/gserver/tests/sequence_nest_layer_group.conf>`_
+* 单层RNN\: `sequence_layer_group.conf <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/legacy/gserver/tests/sequence_layer_group.conf>`_
+* 双层RNN\: `sequence_nest_layer_group.conf <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/legacy/gserver/tests/sequence_nest_layer_group.conf>`_
 
 
 读取双层序列数据
@@ -24,18 +24,18 @@
 
 - 本例中的原始数据一共有10个样本。每个样本由两部分组成，一个label（此处都为2）和一个已经分词后的句子。这个数据也被单层RNN网络直接使用。
 
-..  literalinclude:: ../../../../paddle/gserver/tests/Sequence/tour_train_wdseg
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/Sequence/tour_train_wdseg
     :language: text
 
 
 - 双层序列数据一共有4个样本。 每个样本间用空行分开，整体数据和原始数据完全一样。但于双层序列的LSTM来说，第一个样本同时encode两条数据成两个向量。这四条数据同时处理的句子数量为\ :code:`[2, 3, 2, 3]`\ 。
 
-..  literalinclude:: ../../../../paddle/gserver/tests/Sequence/tour_train_wdseg.nest
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/Sequence/tour_train_wdseg.nest
     :language: text
 
-其次，对于两种不同的输入数据类型，不同DataProvider对比如下(`sequenceGen.py <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/gserver/tests/sequenceGen.py>`_)\：
+其次，对于两种不同的输入数据类型，不同DataProvider对比如下(`sequenceGen.py <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/legacy/gserver/tests/sequenceGen.py>`_)\：
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequenceGen.py
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequenceGen.py
     :language: python
     :lines: 21-39
     :linenos:
@@ -47,7 +47,7 @@
     - words是原始数据中的每一句话，所对应的词表index数组。它是integer_value_sequence类型的，即整数数组。words即为这个数据中的单层时间序列。
     - label是原始数据中对于每一句话的分类标签，它是integer_value类型的。
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequenceGen.py
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequenceGen.py
     :language: python
     :lines: 42-71
     :linenos:
@@ -64,7 +64,7 @@
 
 首先，我们看一下单层RNN的配置。代码中9-15行(高亮部分)即为单层RNN序列的使用代码。这里使用了PaddlePaddle预定义好的RNN处理函数。在这个函数中，RNN对于每一个时间步通过了一个LSTM网络。
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequence_layer_group.conf
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequence_layer_group.conf
     :language: python
     :lines: 38-63
     :linenos:
@@ -85,7 +85,7 @@
 
 * 至此，\ :code:`lstm_last`\ 便和单层RNN配置中的\ :code:`lstm_last`\ 具有相同的结果了。
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequence_nest_layer_group.conf
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequence_nest_layer_group.conf
     :language: python
     :lines: 38-64
     :linenos:
@@ -107,7 +107,7 @@
 
 - 单层RNN：过了一个很简单的recurrent_group。每一个时间步，当前的输入y和上一个时间步的输出rnn_state做了一个全链接。
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequence_rnn.conf
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequence_rnn.conf
     :language: python
     :lines: 36-48
 
@@ -116,7 +116,7 @@
   - 内层inner_step的recurrent_group和单层序列的几乎一样。除了boot_layer=outer_mem，表示将外层的outer_mem作为内层memory的初始状态。外层outer_step中，outer_mem是一个子句的最后一个向量，即整个双层group是将前一个子句的最后一个向量，作为下一个子句memory的初始状态。
   - 从输入数据上看，单双层序列的句子是一样的，只是双层序列将其又做了子序列划分。因此双层序列的配置中，必须将前一个子句的最后一个元素，作为boot_layer传给下一个子句的memory，才能保证和单层序列的配置中“每个时间步都用了上一个时间步的输出结果”一致。
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequence_nest_rnn.conf
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequence_nest_rnn.conf
     :language: python
     :lines: 39-66
 
@@ -134,7 +134,7 @@
 
 **输入不等长** 是指recurrent_group的多个输入序列，在每个时间步的子序列长度可以不相等。但序列输出时，需要指定与某一个输入的序列信息是一致的。使用\ :red:`targetInlink`\ 可以指定哪一个输入和输出序列信息一致，默认指定第一个输入。 
 
-示例3的配置分别为\ `单层不等长RNN <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/gserver/tests/sequence_rnn_multi_unequalength_inputs.py>`_\ 和\ `双层不等长RNN <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/gserver/tests/sequence_nest_rnn_multi_unequalength_inputs.py>`_\ 。
+示例3的配置分别为\ `单层不等长RNN <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/legacy/gserver/tests/sequence_rnn_multi_unequalength_inputs.py>`_\ 和\ `双层不等长RNN <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/legacy/gserver/tests/sequence_nest_rnn_multi_unequalength_inputs.py>`_\ 。
 
 示例3对于单层RNN和双层RNN数据完全相同。
 
@@ -152,14 +152,14 @@
 
 * 单层RNN\:
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequence_rnn_multi_unequalength_inputs.py
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequence_rnn_multi_unequalength_inputs.py
     :language: python
     :lines: 42-59
     :linenos:
 
 * 双层RNN\ \:
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequence_nest_rnn_multi_unequalength_inputs.py
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequence_nest_rnn_multi_unequalength_inputs.py
     :language: python
     :lines: 41-80
     :linenos:

doc/v2/howto/rnn/hrnn_rnn_api_compare_en.rst

@@ -4,7 +4,7 @@
 API comparision between RNN and hierarchical RNN
 #####################
 
-This article takes PaddlePaddle's hierarchical RNN unit test as an example. We will use several examples to illestrate the usage of single-layer and hierarchical RNNs. Each example has two model configurations, one for single-layer, and the other for hierarchical RNN. Although the implementations are different, both the two model configurations' effects are the same. All of the examples in this article only describe the API interface of the hierarchical RNN, while we do not use this hierarchical RNN to solve practical problems. If you want to understand the use of hierarchical RNN in specific issues, please refer to \ :ref:`algo_hrnn_demo`\ 。The unit test file used in this article's example is \ `test_RecurrentGradientMachine.cpp <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/gserver/tests/test_RecurrentGradientMachine.cpp>`_\ 。
+This article takes PaddlePaddle's hierarchical RNN unit test as an example. We will use several examples to illestrate the usage of single-layer and hierarchical RNNs. Each example has two model configurations, one for single-layer, and the other for hierarchical RNN. Although the implementations are different, both the two model configurations' effects are the same. All of the examples in this article only describe the API interface of the hierarchical RNN, while we do not use this hierarchical RNN to solve practical problems. If you want to understand the use of hierarchical RNN in specific issues, please refer to \ :ref:`algo_hrnn_demo`\ 。The unit test file used in this article's example is \ `test_RecurrentGradientMachine.cpp <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/legacy/gserver/tests/test_RecurrentGradientMachine.cpp>`_\ 。
 
 Example 1：Hierarchical RNN without Memory between subsequences
 ================================
@@ -13,8 +13,8 @@ The classical case in the hierarchical RNN is to perform sequence operations on
 
 In this example, the network configuration of single-layer RNNs and hierarchical RNNs are all to use LSTM as en encoder to compress a word-segmented sentence into a vector. The difference is that, RNN uses a hierarchical RNN model, treating multiple sentences as a whole to use encoder to compress simultaneously. They are completely consistent in their semantic meanings. This pair of semantically identical example configurations is as follows：
 
-* RNN\: `sequence_layer_group.conf <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/gserver/tests/sequence_layer_group.conf>`_
-* Hierarchical RNN\: `sequence_nest_layer_group.conf <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/gserver/tests/sequence_nest_layer_group.conf>`_
+* RNN\: `sequence_layer_group.conf <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/legacy/gserver/tests/sequence_layer_group.conf>`_
+* Hierarchical RNN\: `sequence_nest_layer_group.conf <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/legacy/gserver/tests/sequence_nest_layer_group.conf>`_
 
 
 Reading hierarchical sequence data
@@ -24,18 +24,18 @@ Firstly, the original data in this example is as follows \:
 
 - The original data in this example has 10 samples. Each of the sample includes two components: a lable(all 2 here), and a word-segmented sentence. This data is used by single RNN as well. 
 
-..  literalinclude:: ../../../../paddle/gserver/tests/Sequence/tour_train_wdseg
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/Sequence/tour_train_wdseg
     :language: text
 
 
 - The data for hierarchical RNN has 4 samples. Every sample is seperated by a blank line, while the content of the data is the same as the original data. But as for hierarchical LSTM, the first sample will encode two sentences into two vectors simultaneously. The sentence count dealed simultaneously by this 4 samples are \ :code:`[2, 3, 2, 3]`\ .
 
-..  literalinclude:: ../../../../paddle/gserver/tests/Sequence/tour_train_wdseg.nest
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/Sequence/tour_train_wdseg.nest
     :language: text
 
-Secondly, as for these two types of different input data formats, the contrast of different DataProviders are as follows (`sequenceGen.py <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/gserver/tests/sequenceGen.py>`_)\：
+Secondly, as for these two types of different input data formats, the contrast of different DataProviders are as follows (`sequenceGen.py <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/legacy/gserver/tests/sequenceGen.py>`_)\：
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequenceGen.py
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequenceGen.py
     :language: python
     :lines: 21-39
     :linenos:
@@ -47,7 +47,7 @@ Secondly, as for these two types of different input data formats, the contrast o
     - "words" is a list of word table indices corresponding to each word in the sentence in the original data. Its data type is integer_value_sequence, that is integer list. So, "words" is a singler-layer time series in the data. 
     - "label" is the categorical label of each sentence, whose data type is integer_value. 
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequenceGen.py
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequenceGen.py
     :language: python
     :lines: 42-71
     :linenos:
@@ -64,7 +64,7 @@ Model configuration
 
 Firstly, let's look at the configuration of single-layer RNN. The hightlighted part of line 9 to line 15 is the usage of single-layer RNN. Here we use the pre-defined RNN process function in PaddlePaddle. In this function, for each time step, RNN passes through an LSTM network. 
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequence_layer_group.conf
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequence_layer_group.conf
     :language: python
     :lines: 38-63
     :linenos:
@@ -85,7 +85,7 @@ Secondly, let's look at the model configuration of hierarchical RNN which has th
 
 * Till now, \ :code:`lstm_last`\ has the same result as \ :code:`lstm_last`\ in single-layer RNN configuration. 
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequence_nest_layer_group.conf
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequence_nest_layer_group.conf
     :language: python
     :lines: 38-64
     :linenos:
@@ -107,7 +107,7 @@ We select the different parts between single-layer RNN and hierarchical RNN conf
 
 - single-layer RNN：passes through a simple recurrent_group. For each time step, the current input y and the last time step's output rnn_state pass through a fully-connected layer. 
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequence_rnn.conf
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequence_rnn.conf
     :language: python
     :lines: 36-48
 
@@ -116,7 +116,7 @@ We select the different parts between single-layer RNN and hierarchical RNN conf
   - The recurrent_group of inner layer's inner_step is nearly the same as single-layer sequence, except for the case of boot_layer=outer_mem, which means using the outer layer's outer_mem as the initial state for the inner layer's memory. In the outer layer's out_step, outer_mem is the last vector of a subsequence, that is, the whole hierarchical group uses the last vector of the previous subsequence as the initial state for the next subsequence's memory. 
   - From the aspect of the input data, sentences from single-layer and hierarchical RNN are the same. The only difference is that, hierarchical RNN disassembes the sequence into subsequences. So in the hierarchical RNN configuration, we must use the last element of the previous subsequence as a boot_layer for the memory of the next subsequence, so that it makes no difference with "every time step uses the output of last time step" in the sigle-layer RNN configuration. 
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequence_nest_rnn.conf
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequence_nest_rnn.conf
     :language: python
     :lines: 39-66
 
@@ -134,7 +134,7 @@ Example 3：hierarchical RNN with unequal length inputs
 
 **unequal length inputs** means in the multiple input sequences of recurrent_group, the lengths of subsequences can be unequal. But the output of the sequence, needs to be consistent with one of the input sequences. Using \ :red:`targetInlink`\ can help you specify which of the input sequences and the output sequence can be consistent, by default is the first input. 
 
-The configurations of Example 3 are \ `sequence_rnn_multi_unequalength_inputs <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/gserver/tests/sequence_rnn_multi_unequalength_inputs.py>`_ \ and \ `sequence_nest_rnn_multi_unequalength_inputs <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/gserver/tests/sequence_nest_rnn_multi_unequalength_inputs.py>`_\ . 
+The configurations of Example 3 are \ `sequence_rnn_multi_unequalength_inputs <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/legacy/gserver/tests/sequence_rnn_multi_unequalength_inputs.py>`_ \ and \ `sequence_nest_rnn_multi_unequalength_inputs <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/legacy/gserver/tests/sequence_nest_rnn_multi_unequalength_inputs.py>`_\ .
 
 The data for the configurations of Example 3's single-layer RNN and hierarchical RNN are exactly the same. 
 
@@ -152,14 +152,14 @@ Similar to Example 2's configuration, Example 3's configuration uses single-laye
 
 * single-layer RNN\:
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequence_rnn_multi_unequalength_inputs.py
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequence_rnn_multi_unequalength_inputs.py
     :language: python
     :lines: 42-59
     :linenos:
 
 * hierarchical RNN\ \:
 
-..  literalinclude:: ../../../../paddle/gserver/tests/sequence_nest_rnn_multi_unequalength_inputs.py
+..  literalinclude:: ../../../../paddle/legacy/gserver/tests/sequence_nest_rnn_multi_unequalength_inputs.py
     :language: python
     :lines: 41-80
     :linenos:

paddle/CMakeLists.txt

@@ -3,7 +3,7 @@ if(NOT WITH_FLUID_ONLY)
   add_subdirectory(function)
   add_subdirectory(utils)
   add_subdirectory(math)
-  add_subdirectory(gserver)
+  add_subdirectory(legacy/gserver)
   add_subdirectory(parameter)
 
   if(MOBILE_INFERENCE)

paddle/api/GradientMachine.cpp

@@ -16,7 +16,7 @@ limitations under the License. */
 #include "PaddleAPIPrivate.h"
 
 #include "Internal.h"
-#include "paddle/gserver/gradientmachines/NeuralNetwork.h"
+#include "paddle/legacy/gserver/gradientmachines/NeuralNetwork.h"
 
 std::vector<int> GradientMachine::defaultParamTypes = {
     PARAMETER_VALUE, PARAMETER_GRADIENT, PARAMETER_MOMENTUM};

paddle/api/PaddleAPI.h

@@ -19,7 +19,7 @@ limitations under the License. */
 #include <stdexcept>
 #include <string>
 #include <vector>
-#include "paddle/gserver/gradientmachines/GradientMachine.h"
+#include "paddle/legacy/gserver/gradientmachines/GradientMachine.h"
 #include "paddle/utils/Common.h"
 #include "paddle/utils/GlobalConstants.h"
 

paddle/api/PaddleAPIPrivate.h

@@ -14,8 +14,8 @@ limitations under the License. */
 #pragma once
 #include <memory>
 #include "PaddleAPI.h"
-#include "paddle/gserver/evaluators/Evaluator.h"
-#include "paddle/gserver/gradientmachines/GradientMachine.h"
+#include "paddle/legacy/gserver/evaluators/Evaluator.h"
+#include "paddle/legacy/gserver/gradientmachines/GradientMachine.h"
 #include "paddle/parameter/ParameterUpdaterBase.h"
 #include "paddle/trainer/TrainerConfigHelper.h"
 

paddle/api/SequenceGenerator.cpp

@@ -17,7 +17,7 @@ limitations under the License. */
 #include <sstream>
 #include <vector>
 #include "PaddleAPI.h"
-#include "paddle/gserver/gradientmachines/GradientMachine.h"
+#include "paddle/legacy/gserver/gradientmachines/GradientMachine.h"
 #include "paddle/parameter/Argument.h"
 #include "paddle/utils/Flags.h"
 

paddle/api/Trainer.cpp

@@ -19,7 +19,7 @@ limitations under the License. */
 #include <atomic>
 #include <memory>
 
-#include "paddle/gserver/gradientmachines/NeuralNetwork.h"
+#include "paddle/legacy/gserver/gradientmachines/NeuralNetwork.h"
 #include "paddle/trainer/ParamUtil.h"
 #include "paddle/trainer/Trainer.h"
 #include "paddle/trainer/TrainerInternal.h"

paddle/capi/capi_private.h

@@ -13,7 +13,7 @@ See the License for the specific language governing permissions and
 limitations under the License. */
 
 #include "capi.h"
-#include "paddle/gserver/gradientmachines/GradientMachine.h"
+#include "paddle/legacy/gserver/gradientmachines/GradientMachine.h"
 #include "paddle/math/Matrix.h"
 #include "paddle/math/Vector.h"
 #include "paddle/parameter/Argument.h"

paddle/capi/gradient_machine.cpp

@@ -14,7 +14,7 @@ limitations under the License. */
 
 #include "gradient_machine.h"
 #include "capi_private.h"
-#include "paddle/gserver/gradientmachines/NeuralNetwork.h"
+#include "paddle/legacy/gserver/gradientmachines/NeuralNetwork.h"
 
 #define cast(v) paddle::capi::cast<paddle::capi::CGradientMachine>(v)
 

paddle/capi/tests/test_GradientMachine.cpp

@@ -13,7 +13,7 @@ See the License for the specific language governing permissions and
 limitations under the License. */
 
 #include <gtest/gtest.h>
-#include <paddle/gserver/gradientmachines/GradientMachine.h>
+#include <paddle/legacy/gserver/gradientmachines/GradientMachine.h>
 #include <paddle/trainer/TrainerConfigHelper.h>
 #include <stdlib.h>
 #include <string.h>

paddle/gserver/activations/MKLDNNActivation.h → paddle/legacy/gserver/activations/MKLDNNActivation.h

@@ -15,7 +15,7 @@ limitations under the License. */
 #pragma once
 #include "ActivationFunction.h"
 #include "mkldnn.hpp"
-#include "paddle/gserver/layers/MKLDNNBase.h"
+#include "paddle/legacy/gserver/layers/MKLDNNBase.h"
 #include "paddle/math/MKLDNNMatrix.h"
 #include "paddle/parameter/Argument.h"
 

paddle/gserver/evaluators/CTCErrorEvaluator.cpp → paddle/legacy/gserver/evaluators/CTCErrorEvaluator.cpp

@@ -13,7 +13,7 @@ See the License for the specific language governing permissions and
 limitations under the License. */
 
 #include "Evaluator.h"
-#include "paddle/gserver/gradientmachines/NeuralNetwork.h"
+#include "paddle/legacy/gserver/gradientmachines/NeuralNetwork.h"
 #include "paddle/utils/StringUtil.h"
 
 namespace paddle {

paddle/gserver/evaluators/DetectionMAPEvaluator.cpp → paddle/legacy/gserver/evaluators/DetectionMAPEvaluator.cpp

@@ -13,7 +13,7 @@ See the License for the specific language governing permissions and
 limitations under the License. */
 
 #include "Evaluator.h"
-#include "paddle/gserver/layers/DetectionUtil.h"
+#include "paddle/legacy/gserver/layers/DetectionUtil.h"
 
 using std::map;
 using std::vector;

paddle/gserver/evaluators/Evaluator.cpp → paddle/legacy/gserver/evaluators/Evaluator.cpp

@@ -12,8 +12,8 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License. */
 
-#include "paddle/gserver/evaluators/Evaluator.h"
-#include "paddle/gserver/gradientmachines/NeuralNetwork.h"
+#include "paddle/legacy/gserver/evaluators/Evaluator.h"
+#include "paddle/legacy/gserver/gradientmachines/NeuralNetwork.h"
 #include "paddle/utils/Stat.h"
 #include "paddle/utils/StringUtil.h"
 

paddle/gserver/gradientmachines/GradientMachine.h → paddle/legacy/gserver/gradientmachines/GradientMachine.h

@@ -19,15 +19,15 @@ limitations under the License. */
 
 #include "ModelConfig.pb.h"
 #include "TrainerConfig.pb.h"
-#include "paddle/gserver/dataproviders/DataProvider.h"
-#include "paddle/gserver/layers/Layer.h"
+#include "paddle/legacy/gserver/dataproviders/DataProvider.h"
+#include "paddle/legacy/gserver/layers/Layer.h"
 #include "paddle/math/Matrix.h"
 #include "paddle/parameter/Parameter.h"
 #include "paddle/parameter/ParameterUpdaterBase.h"
 #include "paddle/utils/Thread.h"
 
 #ifndef PADDLE_MOBILE_INFERENCE
-#include "paddle/gserver/evaluators/Evaluator.h"
+#include "paddle/legacy/gserver/evaluators/Evaluator.h"
 #endif
 
 namespace paddle {

paddle/gserver/gradientmachines/NeuralNetwork.cpp → paddle/legacy/gserver/gradientmachines/NeuralNetwork.cpp

@@ -21,13 +21,13 @@ limitations under the License. */
 #include "paddle/utils/Stat.h"
 
 #ifdef PADDLE_WITH_MKLDNN
-#include "paddle/gserver/layers/MKLDNNLayer.h"
+#include "paddle/legacy/gserver/layers/MKLDNNLayer.h"
 #endif
 
 #ifndef PADDLE_MOBILE_INFERENCE
 #include "MultiNetwork.h"
 #include "RecurrentGradientMachine.h"
-#include "paddle/gserver/layers/AgentLayer.h"
+#include "paddle/legacy/gserver/layers/AgentLayer.h"
 #endif
 
 namespace paddle {

paddle/gserver/gradientmachines/NeuralNetwork.h → paddle/legacy/gserver/gradientmachines/NeuralNetwork.h

@@ -19,11 +19,11 @@ limitations under the License. */
 #include <memory>
 
 #include "ModelConfig.pb.h"
-#include "paddle/gserver/dataproviders/DataProvider.h"
-#include "paddle/gserver/gradientmachines/GradientMachine.h"
-#include "paddle/gserver/layers/CostLayer.h"
-#include "paddle/gserver/layers/DataLayer.h"
-#include "paddle/gserver/layers/Layer.h"
+#include "paddle/legacy/gserver/dataproviders/DataProvider.h"
+#include "paddle/legacy/gserver/gradientmachines/GradientMachine.h"
+#include "paddle/legacy/gserver/layers/CostLayer.h"
+#include "paddle/legacy/gserver/layers/DataLayer.h"
+#include "paddle/legacy/gserver/layers/Layer.h"
 #include "paddle/parameter/Parameter.h"
 #include "paddle/utils/ClassRegistrar.h"
 

paddle/gserver/gradientmachines/RecurrentGradientMachine.cpp → paddle/legacy/gserver/gradientmachines/RecurrentGradientMachine.cpp

@@ -19,7 +19,7 @@ limitations under the License. */
 #include <functional>
 #include <limits>
 #include "NeuralNetwork.h"
-#include "paddle/gserver/layers/AgentLayer.h"
+#include "paddle/legacy/gserver/layers/AgentLayer.h"
 #include "paddle/utils/Flags.h"
 #include "paddle/utils/Stat.h"
 #include "paddle/utils/Util.h"

paddle/gserver/layers/Layer.h → paddle/legacy/gserver/layers/Layer.h

@@ -18,7 +18,7 @@ limitations under the License. */
 #include <memory>
 #include "ModelConfig.pb.h"
 #include "paddle/function/Function.h"
-#include "paddle/gserver/activations/ActivationFunction.h"
+#include "paddle/legacy/gserver/activations/ActivationFunction.h"
 #include "paddle/math/CpuSparseMatrix.h"
 #include "paddle/parameter/Argument.h"
 #include "paddle/parameter/Parameter.h"

paddle/gserver/layers/RecurrentLayerGroup.cpp → paddle/legacy/gserver/layers/RecurrentLayerGroup.cpp

@@ -13,9 +13,9 @@ See the License for the specific language governing permissions and
 limitations under the License. */
 
 #include <functional>
-#include "paddle/gserver/layers/Layer.h"
+#include "paddle/legacy/gserver/layers/Layer.h"
 
-#include "paddle/gserver/gradientmachines/RecurrentGradientMachine.h"
+#include "paddle/legacy/gserver/gradientmachines/RecurrentGradientMachine.h"
 #include "paddle/utils/Stat.h"
 
 namespace paddle {

paddle/gserver/layers/ValidationLayer.h → paddle/legacy/gserver/layers/ValidationLayer.h

@@ -16,7 +16,7 @@ limitations under the License. */
 #include <memory>
 
 #include "Layer.h"
-#include "paddle/gserver/evaluators/Evaluator.h"
+#include "paddle/legacy/gserver/evaluators/Evaluator.h"
 
 DECLARE_int32(trainer_id);
 

paddle/gserver/tests/CMakeLists.txt → paddle/legacy/gserver/tests/CMakeLists.txt

@@ -36,7 +36,7 @@ gserver_test(test_Upsample)
 
 set(PYTHON_PATH 
    ${PADDLE_SOURCE_DIR}/paddle/.set_python_path.sh -d 
-   ${PADDLE_BINARY_DIR}/python/:${PADDLE_BINARY_DIR}/paddle/gserver/tests)
+   ${PADDLE_BINARY_DIR}/python/:${PADDLE_BINARY_DIR}/paddle/legacy/gserver/tests)
 function(gserver_test_with_python TARGET)
   add_unittest_without_exec(${TARGET} ${TARGET}.cpp)
   add_test(NAME ${TARGET}

paddle/gserver/tests/LayerGradUtil.h → paddle/legacy/gserver/tests/LayerGradUtil.h

@@ -14,7 +14,7 @@ limitations under the License. */
 
 #pragma once
 #include "ModelConfig.pb.h"
-#include "paddle/gserver/layers/DataLayer.h"
+#include "paddle/legacy/gserver/layers/DataLayer.h"
 
 #include "paddle/testing/TestUtil.h"
 using namespace std;  // NOLINT

paddle/gserver/tests/MKLDNNTester.cpp → paddle/legacy/gserver/tests/MKLDNNTester.cpp

@@ -13,8 +13,8 @@ See the License for the specific language governing permissions and
 limitations under the License. */
 
 #include "MKLDNNTester.h"
-#include "paddle/gserver/layers/MKLDNNBase.h"
-#include "paddle/gserver/layers/MKLDNNLayer.h"
+#include "paddle/legacy/gserver/layers/MKLDNNBase.h"
+#include "paddle/legacy/gserver/layers/MKLDNNLayer.h"
 #include "paddle/trainer/Trainer.h"
 
 namespace paddle {

paddle/gserver/tests/MKLDNNTester.h → paddle/legacy/gserver/tests/MKLDNNTester.h

@@ -17,8 +17,8 @@ limitations under the License. */
 #include <string>
 #include <vector>
 #include "LayerGradUtil.h"
-#include "paddle/gserver/layers/MKLDNNBase.h"
-#include "paddle/gserver/layers/MKLDNNLayer.h"
+#include "paddle/legacy/gserver/layers/MKLDNNBase.h"
+#include "paddle/legacy/gserver/layers/MKLDNNLayer.h"
 
 namespace paddle {
 

paddle/gserver/tests/test_ActivationGrad.cpp → paddle/legacy/gserver/tests/test_ActivationGrad.cpp

@@ -16,7 +16,7 @@ limitations under the License. */
 #include <string>
 #include <vector>
 #include "ModelConfig.pb.h"
-#include "paddle/gserver/layers/DataLayer.h"
+#include "paddle/legacy/gserver/layers/DataLayer.h"
 
 #include "LayerGradUtil.h"
 #include "paddle/testing/TestUtil.h"

paddle/gserver/tests/test_BatchNorm.cpp → paddle/legacy/gserver/tests/test_BatchNorm.cpp

@@ -16,7 +16,7 @@ limitations under the License. */
 #include <string>
 #include <vector>
 #include "ModelConfig.pb.h"
-#include "paddle/gserver/layers/DataLayer.h"
+#include "paddle/legacy/gserver/layers/DataLayer.h"
 #include "paddle/utils/GlobalConstants.h"
 
 #include "LayerGradUtil.h"

paddle/gserver/tests/test_CRFLayerGrad.cpp → paddle/legacy/gserver/tests/test_CRFLayerGrad.cpp

@@ -14,8 +14,8 @@ limitations under the License. */
 
 #include <gtest/gtest.h>
 #include "ModelConfig.pb.h"
-#include "paddle/gserver/layers/DataLayer.h"
-#include "paddle/gserver/layers/LinearChainCRF.h"
+#include "paddle/legacy/gserver/layers/DataLayer.h"
+#include "paddle/legacy/gserver/layers/LinearChainCRF.h"
 
 #include "LayerGradUtil.h"
 #include "paddle/testing/TestUtil.h"

paddle/gserver/tests/test_ConvTrans.cpp → paddle/legacy/gserver/tests/test_ConvTrans.cpp

@@ -16,7 +16,7 @@ limitations under the License. */
 #include <string>
 #include <vector>
 #include "ModelConfig.pb.h"
-#include "paddle/gserver/layers/DataLayer.h"
+#include "paddle/legacy/gserver/layers/DataLayer.h"
 #include "paddle/math/MathUtils.h"
 #include "paddle/utils/GlobalConstants.h"
 

paddle/gserver/tests/test_ConvUnify.cpp → paddle/legacy/gserver/tests/test_ConvUnify.cpp

@@ -16,7 +16,7 @@ limitations under the License. */
 #include <string>
 #include <vector>
 #include "ModelConfig.pb.h"
-#include "paddle/gserver/layers/DataLayer.h"
+#include "paddle/legacy/gserver/layers/DataLayer.h"
 #include "paddle/math/MathUtils.h"
 #include "paddle/utils/GlobalConstants.h"
 

paddle/gserver/tests/test_CrossEntropyOverBeamGrad.cpp → paddle/legacy/gserver/tests/test_CrossEntropyOverBeamGrad.cpp

@@ -17,7 +17,7 @@ limitations under the License. */
 
 #include <gtest/gtest.h>
 #include "ModelConfig.pb.h"
-#include "paddle/gserver/layers/DataLayer.h"
+#include "paddle/legacy/gserver/layers/DataLayer.h"
 
 #include "LayerGradUtil.h"
 #include "paddle/testing/TestUtil.h"

paddle/gserver/tests/test_KmaxSeqScore.cpp → paddle/legacy/gserver/tests/test_KmaxSeqScore.cpp

@@ -17,7 +17,7 @@ limitations under the License. */
 #include <string>
 #include <vector>
 #include "ModelConfig.pb.h"
-#include "paddle/gserver/layers/DataLayer.h"
+#include "paddle/legacy/gserver/layers/DataLayer.h"
 #include "paddle/utils/GlobalConstants.h"
 
 #include "LayerGradUtil.h"

paddle/gserver/tests/test_LayerGrad.cpp → paddle/legacy/gserver/tests/test_LayerGrad.cpp

@@ -19,7 +19,7 @@ limitations under the License. */
 #include <string>
 #include <vector>
 #include "ModelConfig.pb.h"
-#include "paddle/gserver/layers/DataLayer.h"
+#include "paddle/legacy/gserver/layers/DataLayer.h"
 #include "paddle/math/MathUtils.h"
 
 #include "LayerGradUtil.h"

paddle/gserver/tests/test_LinearChainCRF.cpp → paddle/legacy/gserver/tests/test_LinearChainCRF.cpp

@@ -14,7 +14,7 @@ limitations under the License. */
 
 #include <gtest/gtest.h>
 #include <vector>
-#include "paddle/gserver/layers/LinearChainCRF.h"
+#include "paddle/legacy/gserver/layers/LinearChainCRF.h"
 #include "paddle/utils/Util.h"
 
 using namespace paddle;  // NOLINT

paddle/gserver/tests/test_MKLDNN.cpp → paddle/legacy/gserver/tests/test_MKLDNN.cpp

@@ -18,7 +18,7 @@ limitations under the License. */
 #include <vector>
 #include "MKLDNNTester.h"
 #include "ModelConfig.pb.h"
-#include "paddle/gserver/activations/MKLDNNActivation.h"
+#include "paddle/legacy/gserver/activations/MKLDNNActivation.h"
 #include "paddle/math/MathUtils.h"
 
 using namespace paddle;  // NOLINT

paddle/gserver/tests/test_MultinomialSampler.cpp → paddle/legacy/gserver/tests/test_MultinomialSampler.cpp

@@ -20,7 +20,7 @@ limitations under the License. */
 #undef PADDLE_DISABLE_TIMER
 #include "paddle/utils/Stat.h"
 
-#include "paddle/gserver/layers/MultinomialSampler.h"
+#include "paddle/legacy/gserver/layers/MultinomialSampler.h"
 #include "paddle/utils/Util.h"
 
 using namespace paddle;  // NOLINT

paddle/gserver/tests/test_PyDataProvider.cpp → paddle/legacy/gserver/tests/test_PyDataProvider.cpp

@@ -17,7 +17,7 @@ limitations under the License. */
 
 #include <gtest/gtest.h>
 
-#include "paddle/gserver/dataproviders/PyDataProvider.h"
+#include "paddle/legacy/gserver/dataproviders/PyDataProvider.h"
 #include "paddle/utils/Util.h"
 
 #include "paddle/testing/TestUtil.h"

paddle/gserver/tests/test_PyDataProvider2.cpp → paddle/legacy/gserver/tests/test_PyDataProvider2.cpp

@@ -15,7 +15,7 @@ limitations under the License. */
 #ifndef PADDLE_NO_PYTHON
 #include <gtest/gtest.h>
 #include <fstream>
-#include "paddle/gserver/dataproviders/DataProvider.h"
+#include "paddle/legacy/gserver/dataproviders/DataProvider.h"
 #include "paddle/utils/PythonUtil.h"
 #include "paddle/utils/Util.h"
 

paddle/gserver/tests/test_RecurrentGradientMachine.cpp → paddle/legacy/gserver/tests/test_RecurrentGradientMachine.cpp

@@ -13,7 +13,7 @@ See the License for the specific language governing permissions and
 limitations under the License. */
 
 #include <gtest/gtest.h>
-#include <paddle/gserver/gradientmachines/GradientMachine.h>
+#include <paddle/legacy/gserver/gradientmachines/GradientMachine.h>
 #include <paddle/parameter/ParameterUpdateFunctions.h>
 #include <paddle/trainer/Trainer.h>
 #include <paddle/trainer/TrainerInternal.h>

paddle/gserver/tests/test_RecurrentLayer.cpp → paddle/legacy/gserver/tests/test_RecurrentLayer.cpp

@@ -16,8 +16,8 @@ limitations under the License. */
 #include <paddle/utils/Version.h>
 #include <vector>
 #include "ModelConfig.pb.h"
-#include "paddle/gserver/layers/DataLayer.h"
-#include "paddle/gserver/layers/Layer.h"
+#include "paddle/legacy/gserver/layers/DataLayer.h"
+#include "paddle/legacy/gserver/layers/Layer.h"
 
 #include "paddle/testing/TestUtil.h"
 
@@ -220,9 +220,9 @@ TEST(Layer, RecurrentLayer) {
 }
 
 #define protected public
-#include "paddle/gserver/layers/GatedRecurrentLayer.h"
-#include "paddle/gserver/layers/LstmLayer.h"
-#include "paddle/gserver/layers/RecurrentLayer.h"
+#include "paddle/legacy/gserver/layers/GatedRecurrentLayer.h"
+#include "paddle/legacy/gserver/layers/LstmLayer.h"
+#include "paddle/legacy/gserver/layers/RecurrentLayer.h"
 template <class T>
 class TestRecurrentLayer {
  public:
@@ -423,7 +423,7 @@ TEST(Layer, LstmLayer) {
 
 #ifdef PADDLE_WITH_MKLML
 
-#include "paddle/gserver/layers/MKLPackedRecurrentLayer.h"
+#include "paddle/legacy/gserver/layers/MKLPackedRecurrentLayer.h"
 
 LayerPtr initMKLPackedLayer(LayerConfig layerConfig,
                             bool reversed,

paddle/gserver/tests/test_SelectiveFCLayer.cpp → paddle/legacy/gserver/tests/test_SelectiveFCLayer.cpp

@@ -19,10 +19,10 @@ limitations under the License. */
 #include <cstdlib>
 #include <ctime>
 #include "ModelConfig.pb.h"
-#include "paddle/gserver/layers/DataLayer.h"
-#include "paddle/gserver/layers/FullyConnectedLayer.h"
-#include "paddle/gserver/layers/Layer.h"
-#include "paddle/gserver/layers/SelectiveFullyConnectedLayer.h"
+#include "paddle/legacy/gserver/layers/DataLayer.h"
+#include "paddle/legacy/gserver/layers/FullyConnectedLayer.h"
+#include "paddle/legacy/gserver/layers/Layer.h"
+#include "paddle/legacy/gserver/layers/SelectiveFullyConnectedLayer.h"
 #include "paddle/math/CpuSparseMatrix.h"
 
 using namespace paddle;  // NOLINT

paddle/gserver/tests/test_SeqSliceLayerGrad.cpp → paddle/legacy/gserver/tests/test_SeqSliceLayerGrad.cpp

@@ -14,7 +14,7 @@ limitations under the License. */
 
 #include <gtest/gtest.h>
 #include "ModelConfig.pb.h"
-#include "paddle/gserver/layers/DataLayer.h"
+#include "paddle/legacy/gserver/layers/DataLayer.h"
 
 #include "LayerGradUtil.h"
 #include "paddle/testing/TestUtil.h"

paddle/gserver/tests/test_WarpCTCLayer.cpp → paddle/legacy/gserver/tests/test_WarpCTCLayer.cpp

@@ -15,10 +15,10 @@ limitations under the License. */
 #include <gtest/gtest.h>
 #include <paddle/utils/Version.h>
 #include "ModelConfig.pb.h"
-#include "paddle/gserver/layers/CTCLayer.h"
-#include "paddle/gserver/layers/DataLayer.h"
-#include "paddle/gserver/layers/Layer.h"
-#include "paddle/gserver/layers/WarpCTCLayer.h"
+#include "paddle/legacy/gserver/layers/CTCLayer.h"
+#include "paddle/legacy/gserver/layers/DataLayer.h"
+#include "paddle/legacy/gserver/layers/Layer.h"
+#include "paddle/legacy/gserver/layers/WarpCTCLayer.h"
 
 #include "paddle/testing/TestUtil.h"
 

paddle/trainer/ParamUtil.cpp

@@ -31,8 +31,8 @@ limitations under the License. */
 #include "paddle/utils/Util.h"
 
 #include "TesterConfig.h"
-#include "paddle/gserver/gradientmachines/NeuralNetwork.h"
-#include "paddle/gserver/layers/ValidationLayer.h"
+#include "paddle/legacy/gserver/gradientmachines/NeuralNetwork.h"
+#include "paddle/legacy/gserver/layers/ValidationLayer.h"
 
 namespace paddle {
 

paddle/trainer/ParamUtil.h

@@ -19,8 +19,8 @@ limitations under the License. */
 #include <stdio.h>
 
 #include "hl_gpu.h"
-#include "paddle/gserver/dataproviders/DataProvider.h"
-#include "paddle/gserver/gradientmachines/GradientMachine.h"
+#include "paddle/legacy/gserver/dataproviders/DataProvider.h"
+#include "paddle/legacy/gserver/gradientmachines/GradientMachine.h"
 
 #include <stdlib.h>
 #include <fstream>

paddle/trainer/ParameterUpdater.h

@@ -25,7 +25,7 @@ limitations under the License. */
 #include "paddle/parameter/ParameterUpdaterBase.h"
 
 #include "TrainerConfig.pb.h"
-#include "paddle/gserver/layers/Layer.h"
+#include "paddle/legacy/gserver/layers/Layer.h"
 
 #include <memory>
 #include <vector>

paddle/trainer/Tester.cpp

@@ -30,9 +30,9 @@ limitations under the License. */
 #include "paddle/utils/Util.h"
 
 #include "TesterConfig.h"
-#include "paddle/gserver/gradientmachines/GradientMachineMode.h"
-#include "paddle/gserver/gradientmachines/NeuralNetwork.h"
-#include "paddle/gserver/layers/ValidationLayer.h"
+#include "paddle/legacy/gserver/gradientmachines/GradientMachineMode.h"
+#include "paddle/legacy/gserver/gradientmachines/NeuralNetwork.h"
+#include "paddle/legacy/gserver/layers/ValidationLayer.h"
 
 namespace paddle {
 

paddle/trainer/Tester.h

@@ -19,8 +19,8 @@ limitations under the License. */
 #include <stdio.h>
 
 #include "hl_gpu.h"
-#include "paddle/gserver/dataproviders/DataProvider.h"
-#include "paddle/gserver/gradientmachines/GradientMachine.h"
+#include "paddle/legacy/gserver/dataproviders/DataProvider.h"
+#include "paddle/legacy/gserver/gradientmachines/GradientMachine.h"
 
 #include "TrainerConfig.pb.h"
 

paddle/trainer/TesterConfig.h

@@ -19,7 +19,7 @@ limitations under the License. */
 #include <stdio.h>
 
 #include "hl_gpu.h"
-#include "paddle/gserver/gradientmachines/GradientMachine.h"
+#include "paddle/legacy/gserver/gradientmachines/GradientMachine.h"
 
 #include "TrainerConfig.pb.h"
 

paddle/trainer/Trainer.cpp

@@ -33,9 +33,9 @@ limitations under the License. */
 #include "TesterConfig.h"
 #include "ThreadParameterUpdater.h"
 #include "TrainerConfigHelper.h"
-#include "paddle/gserver/gradientmachines/GradientMachineMode.h"
-#include "paddle/gserver/gradientmachines/NeuralNetwork.h"
-#include "paddle/gserver/layers/ValidationLayer.h"
+#include "paddle/legacy/gserver/gradientmachines/GradientMachineMode.h"
+#include "paddle/legacy/gserver/gradientmachines/NeuralNetwork.h"
+#include "paddle/legacy/gserver/layers/ValidationLayer.h"
 
 DEFINE_string(config, "", "Trainer config file");
 

paddle/trainer/Trainer.h

@@ -19,8 +19,8 @@ limitations under the License. */
 #include <stdio.h>
 
 #include "hl_gpu.h"
-#include "paddle/gserver/dataproviders/DataProvider.h"
-#include "paddle/gserver/gradientmachines/GradientMachine.h"
+#include "paddle/legacy/gserver/dataproviders/DataProvider.h"
+#include "paddle/legacy/gserver/gradientmachines/GradientMachine.h"
 
 #include <stdlib.h>
 #include <fstream>

paddle/trainer/TrainerInternal.cpp

@@ -24,8 +24,8 @@ limitations under the License. */
 
 #include <google/protobuf/text_format.h>
 
-#include "paddle/gserver/gradientmachines/NeuralNetwork.h"
-#include "paddle/gserver/layers/ValidationLayer.h"
+#include "paddle/legacy/gserver/gradientmachines/NeuralNetwork.h"
+#include "paddle/legacy/gserver/layers/ValidationLayer.h"
 #include "paddle/utils/GlobalConstants.h"
 #include "paddle/utils/PythonUtil.h"
 #include "paddle/utils/Stat.h"

paddle/trainer/TrainerInternal.h

@@ -25,7 +25,7 @@ limitations under the License. */
 #include "TrainerConfigHelper.h"
 #include "TrainerInternalConfig.h"
 #include "hl_gpu.h"
-#include "paddle/gserver/gradientmachines/GradientMachine.h"
+#include "paddle/legacy/gserver/gradientmachines/GradientMachine.h"
 
 namespace paddle {
 

paddle/trainer/TrainerInternalConfig.h

@@ -19,7 +19,7 @@ limitations under the License. */
 #include <stdio.h>
 
 #include "hl_gpu.h"
-#include "paddle/gserver/gradientmachines/GradientMachine.h"
+#include "paddle/legacy/gserver/gradientmachines/GradientMachine.h"
 
 #include "TrainerConfig.pb.h"
 

paddle/trainer/tests/test_PyDataProviderWrapper.cpp

@@ -15,7 +15,7 @@ limitations under the License. */
 #ifndef PADDLE_NO_PYTHON
 #include <DataConfig.pb.h>
 #include <gtest/gtest.h>
-#include <paddle/gserver/dataproviders/DataProvider.h>
+#include <paddle/legacy/gserver/dataproviders/DataProvider.h>
 #include <paddle/math/Matrix.h>
 #include <paddle/parameter/Argument.h>
 #include <paddle/utils/PythonUtil.h>

python/paddle/trainer_config_helpers/layers.py

@@ -4182,9 +4182,9 @@ def recurrent_group(step, input, reverse=False, name=None, targetInlink=None):
 
     You can see following configs for further usages:
 
-    - time steps: lstmemory_group, paddle/gserver/tests/sequence_layer_group.conf, \
+    - time steps: lstmemory_group, paddle/legacy/gserver/tests/sequence_layer_group.conf, \
                   demo/seqToseq/seqToseq_net.py
-    - sequence steps: paddle/gserver/tests/sequence_nest_layer_group.conf
+    - sequence steps: paddle/legacy/gserver/tests/sequence_nest_layer_group.conf
 
     :param step: A step function which takes the input of recurrent_group as its own
                  input and returns values as recurrent_group's output every time step.

tools/codestyle/cpplint_pre_commit.hook

@@ -4,7 +4,7 @@ TOTAL_ERRORS=0
 
 # The trick to remove deleted files: https://stackoverflow.com/a/2413151
 for file in $(git diff --cached --name-status | awk '$1 != "D" {print $2}'); do
-    if [[ $file =~ ^(paddle/api/.*|paddle/capi/.*|paddle/contrib/.*|paddle/cuda/.*|paddle/function/.*|paddle/gserver/.*|paddle/math/.*|paddle/optimizer/.*|paddle/parameter/.*|paddle/pserver/.*|paddle/trainer/.*|paddle/utils/.*) ]]; then
+    if [[ $file =~ ^(paddle/api/.*|paddle/capi/.*|paddle/contrib/.*|paddle/cuda/.*|paddle/function/.*|paddle/legacy/gserver/.*|paddle/math/.*|paddle/optimizer/.*|paddle/parameter/.*|paddle/pserver/.*|paddle/trainer/.*|paddle/utils/.*) ]]; then
         continue;
     else
         cpplint --filter=-readability/fn_size $file;