merge baidu/develop

.gitignore

@@ -27,7 +27,6 @@ third_party/
 cmake-build-*
 
 # generated while compiling
-python/paddle/v2/fluid/core.so
 paddle/pybind/pybind.h
 CMakeFiles
 cmake_install.cmake

doc/api/fluid/index.rst

+======================
+Fluid
+======================
+
+..  toctree::
+    :maxdepth: 1
+
+    layers.rst
+    data_feeder.rst
+    executor.rst
+    initializer.rst
+    evaluator.rst
+    nets.rst
+    optimizer.rst
+    param_attr.rst
+    profiler.rst
+    regularizer.rst
+    io.rst

doc/api/index_en.rst

@@ -8,4 +8,4 @@ API
     v2/model_configs.rst
     v2/data.rst
     v2/run_logic.rst
-    v2/fluid.rst
+    fluid/index.rst

doc/api/v2/fluid.rst

-======================
-Fluid
-======================
-
-..  toctree::
-    :maxdepth: 1
-
-    fluid/layers.rst
-    fluid/data_feeder.rst
-    fluid/executor.rst
-    fluid/initializer.rst
-    fluid/evaluator.rst
-    fluid/nets.rst
-    fluid/optimizer.rst
-    fluid/param_attr.rst
-    fluid/profiler.rst
-    fluid/regularizer.rst
-    fluid/io.rst

doc/design/concurrent_programming.md

@@ -12,7 +12,7 @@ The following table compares concepts in Fluid and Go
 
 | Go | Fluid |
 |----|-------|
-|user-defined functions | [layers](https://github.com/PaddlePaddle/Paddle/tree/develop/python/paddle/v2/fluid) |
+|user-defined functions | [layers](https://github.com/PaddlePaddle/Paddle/tree/develop/python/paddle/fluid) |
 | control-flow and built-in functions | [intrinsics/operators](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/operators) |
 | goroutines, channels | [class ThreadPool](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/framework/thread_pool.h) |
 | runtime | [class Executor](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/executor.h) |

doc/design/fluid.md

@@ -89,7 +89,7 @@ with train_loop.block():
     h[t] = the_step(input[t])
 ```    
 
-An actual Fluid example is described  [here](https://github.com/PaddlePaddle/Paddle/blob/a91efdde6910ce92a78e3aa7157412c4c88d9ee8/python/paddle/v2/fluid/tests/test_while_op.py#L36-L44).
+An actual Fluid example is described  [here](https://github.com/PaddlePaddle/Paddle/blob/bde090a97564b9c61a6aaa38b72ccc4889d102d9/python/paddle/fluid/tests/unittests/test_while_op.py#L50-L58).
 
 From the example, the Fluid programs look very similar to their PyTorch equivalent programs, except that Fluid's loop structure, wrapped with Python's `with` statement, could run much faster than just a Python loop.
 

doc/design/memory_optimization.md

@@ -101,7 +101,7 @@ In-place is a built-in attribute of an operator. Since we treat in-place and oth
 
 #### contruct control flow graph
 
-Following is the ProgramDesc protobuf of [machine translation](https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/v2/fluid/tests/book/test_machine_translation.py) example.
+Following is the ProgramDesc protobuf of [machine translation](https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/fluid/tests/book/test_machine_translation.py) example.
 
 - Block0:
 

doc/howto/capi/index_cn.rst

 C-API预测库
 ==================
 
+当我们训练完一个神经网络模型之后，下一步就是用模型来做预测。预测就是准备输入数据，经过模型处理之后，得到预测结果的过程。
+
+相比于模型训练，预测有如下特点：
+
+#. 预测不需要训练过程中反向传播和参数更新的部分。
+#. 预测不需要标签(label)。
+#. 预测很多时候需要和用户系统整合在一起。
+
+因为上述特点，模型预测SDK需要单独设计，并具备以下特点：
+
+#. 预测SDK不包含反向传播和参数更新部分，以减小SDK的体积。
+#. 预测SDK需要提供一个简洁的用户接口，方便使用。
+#. 因为输入数据可能有多种结构，对输入数据的格式做清晰简洁的封装。
+#. 为了和用户系统兼容，SDK的接口需要是满足C标准的接口。
+
+PaddlePaddle提供了C-API，用于解决上述问题。关于C-API的使用，我们提供了如下指南：
+
 ..  toctree::
   :maxdepth: 1
 

doc/howto/cluster/fluid_cluster_train_en.md

@@ -32,7 +32,7 @@ The non-cluster version of this demo with fluid API is as follows:
 
 ``` python
 import paddle.v2 as paddle
-import paddle.v2.fluid as fluid
+import paddle.fluid as fluid
 
 x = fluid.layers.data(name='x', shape=[13], dtype='float32')
 y_predict = fluid.layers.fc(input=x, size=1, act=None)
@@ -125,11 +125,11 @@ for pass_id in range(100):
 
 ### E2E demo
 
-Please find the complete demo from [here](https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/v2/fluid/tests/book_distribute/notest_dist_fit_a_line.py).
+Please find the complete demo from [here](https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/fluid/tests/book_distribute/notest_dist_fit_a_line.py).
 First `cd` into the folder that contains the `python` files. In this case:
 
 ```bash
-cd /paddle/python/paddle/v2/fluid/tests/book_distribute
+cd /paddle/python/paddle/fluid/tests/book_distribute
 ```
 
 In parameter server node run the following in the command line:

doc/howto/optimization/cpu_profiling_cn.md

@@ -35,7 +35,7 @@ cprofilev -a 0.0.0.0 -p 3214 -f profile.out main.py
 ```
    ncalls  tottime  percall  cumtime  percall filename:lineno(function)
         1    0.284    0.284   29.514   29.514 main.py:1(<module>)
-     4696    0.128    0.000   15.748    0.003 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/executor.py:20(run)
+     4696    0.128    0.000   15.748    0.003 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/executor.py:20(run)
      4696   12.040    0.003   12.040    0.003 {built-in method run}
         1    0.144    0.144    6.534    6.534 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/__init__.py:14(<module>)
 ```
@@ -61,9 +61,9 @@ cprofilev -a 0.0.0.0 -p 3214 -f profile.out main.py
 ```text
      4696   12.040    0.003   12.040    0.003 {built-in method run}
    300005    0.874    0.000    1.681    0.000 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/dataset/mnist.py:38(reader)
-   107991    0.676    0.000    1.519    0.000 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:219(__init__)
-     4697    0.626    0.000    2.291    0.000 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:428(sync_with_cpp)
-        1    0.618    0.618    0.618    0.618 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/__init__.py:1(<module>)
+   107991    0.676    0.000    1.519    0.000 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/framework.py:219(__init__)
+     4697    0.626    0.000    2.291    0.000 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/framework.py:428(sync_with_cpp)
+        1    0.618    0.618    0.618    0.618 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/__init__.py:1(<module>)
 ```
 
 可以看到最耗时的函数是C++端的`run`函数。这需要联合我们第二节`Python`与`C++`混合代码的性能分析来进行调优。而`sync_with_cpp`函数的总共耗时很长，每次调用的耗时也很长。于是我们可以点击`sync_with_cpp`的详细信息，了解其调用关系。
@@ -76,9 +76,9 @@ Called By:
 
 Function                                                                                                 was called by...
                                                                                                              ncalls  tottime  cumtime
-/home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:428(sync_with_cpp)  <-    4697    0.626    2.291  /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:562(sync_with_cpp)
-/home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:562(sync_with_cpp)  <-    4696    0.019    2.316  /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:487(clone)
-                                                                                                                  1    0.000    0.001  /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:534(append_backward)
+/home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/framework.py:428(sync_with_cpp)  <-    4697    0.626    2.291  /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/framework.py:562(sync_with_cpp)
+/home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/framework.py:562(sync_with_cpp)  <-    4696    0.019    2.316  /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/framework.py:487(clone)
+                                                                                                                  1    0.000    0.001  /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/framework.py:534(append_backward)
 
 
 Called:

doc/howto/optimization/cpu_profiling_en.md

@@ -49,7 +49,7 @@ port, we will see the output like the following:
 ```
    ncalls  tottime  percall  cumtime  percall filename:lineno(function)
         1    0.284    0.284   29.514   29.514 main.py:1(<module>)
-     4696    0.128    0.000   15.748    0.003 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/executor.py:20(run)
+     4696    0.128    0.000   15.748    0.003 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/executor.py:20(run)
      4696   12.040    0.003   12.040    0.003 {built-in method run}
         1    0.144    0.144    6.534    6.534 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/__init__.py:14(<module>)
 ```
@@ -74,9 +74,9 @@ focus on. We can sort above profiling file by tottime:
 ```text
      4696   12.040    0.003   12.040    0.003 {built-in method run}
    300005    0.874    0.000    1.681    0.000 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/dataset/mnist.py:38(reader)
-   107991    0.676    0.000    1.519    0.000 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:219(__init__)
-     4697    0.626    0.000    2.291    0.000 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:428(sync_with_cpp)
-        1    0.618    0.618    0.618    0.618 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/__init__.py:1(<module>)
+   107991    0.676    0.000    1.519    0.000 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/framework.py:219(__init__)
+     4697    0.626    0.000    2.291    0.000 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/framework.py:428(sync_with_cpp)
+        1    0.618    0.618    0.618    0.618 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/__init__.py:1(<module>)
 ```
 
 We can see that the most time-consuming function is the `built-in
@@ -93,9 +93,9 @@ Called By:
 
 Function                                                                                                 was called by...
                                                                                                              ncalls  tottime  cumtime
-/home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:428(sync_with_cpp)  <-    4697    0.626    2.291  /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:562(sync_with_cpp)
-/home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:562(sync_with_cpp)  <-    4696    0.019    2.316  /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:487(clone)
-                                                                                                                  1    0.000    0.001  /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:534(append_backward)
+/home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/framework.py:428(sync_with_cpp)  <-    4697    0.626    2.291  /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/framework.py:562(sync_with_cpp)
+/home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/framework.py:562(sync_with_cpp)  <-    4696    0.019    2.316  /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/framework.py:487(clone)
+                                                                                                                  1    0.000    0.001  /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/fluid/framework.py:534(append_backward)
 
 
 Called:

doc/howto/read_source.md

 # PaddlePaddle Fluid Source Code Overview
 
-Examples: https://github.com/PaddlePaddle/Paddle/tree/develop/python/paddle/v2/fluid/tests/book
+Examples: https://github.com/PaddlePaddle/Paddle/tree/develop/python/paddle/fluid/tests/book
 
 Core: https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/framework
 
@@ -26,16 +26,16 @@ sgd_optimizer = fluid.optimizer.SGD(learning_rate=0.001)
 sgd_optimizer.minimize(avg_cost)
 ```
 
-- Variables: `x`,  `y`, `y_predict`, `cost` and `avg_cost`. [Python](https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/v2/fluid/framework.py#)
-- Layers: `fluid.layers.data`, `fluid.layers.fc` and `fluid.layers.mean` are layers. [Python](https://github.com/PaddlePaddle/Paddle/tree/develop/python/paddle/v2/fluid/layers)
+- Variables: `x`,  `y`, `y_predict`, `cost` and `avg_cost`. [Python](https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/fluid/framework.py#)
+- Layers: `fluid.layers.data`, `fluid.layers.fc` and `fluid.layers.mean` are layers. [Python](https://github.com/PaddlePaddle/Paddle/tree/develop/python/paddle/fluid/layers)
   - Every Layer has one or more operators and variables/parameters
     - All the operators are defined at [`paddle/operators/`](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/operators). Other worth-looking files:
       - Base class: [`paddle/framework/operator.h`](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/operator.h)
       - Operator Registration: [`paddle/framework/op_registry.h`](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/op_registry.h) 
       - Operator Lookup: [`paddle/framework/op_info.h`](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/op_info.h)
 - Optimizer: `fluid.optimizer.SGD`. It does the following
-  - Add backward operators. [[Python](https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/v2/fluid/backward.py)]
-  - Add optimizer operators. [[Python](https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/v2/fluid/optimizer.py)]
+  - Add backward operators. [[Python](https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/fluid/backward.py)]
+  - Add optimizer operators. [[Python](https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/fluid/optimizer.py)]
 
 # Run Time
 
@@ -57,7 +57,7 @@ exe.run(fluid.default_main_program(),
 
 - Place: `place`. one of CPU, GPU or FPGA. [C++](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/platform/place.h)
   - The device handle are at [paddle/platform/device_context.h](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/platform/device_context.h)
-- Executor: `fluid.Executor(place)`. [[Python](https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/v2/fluid/executor.py), [C++](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/executor.cc)]
+- Executor: `fluid.Executor(place)`. [[Python](https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/fluid/executor.py), [C++](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/executor.cc)]
   - Feeds the data: `feed=feeder.feed(data)`
   - Evaluates all the operators
   - Fetches the result: `fetch_list=[avg_cost]`

paddle/fluid/framework/executor.cc

@@ -58,13 +58,13 @@ static void CreateTensor(Variable* var, proto::VarType::Type var_type) {
     var->GetMutable<ReaderHolder>();
   } else if (var_type == proto::VarType::CHANNEL) {
     var->GetMutable<ChannelHolder>();
-  } else if (var_type == proto::VarType::NCCL_COM) {
-    // GetMutable will be called in ncclInit
+  } else if (var_type == proto::VarType::RAW) {
+    // GetMutable will be called in operator
   } else {
     PADDLE_THROW(
         "Variable type %d is not in "
         "[LOD_TENSOR, SELECTED_ROWS, FEED_MINIBATCH, FETCH_LIST, "
-        "LOD_RANK_TABLE, PLACE_LIST, READER, CHANNEL, NCCL_COM]",
+        "LOD_RANK_TABLE, PLACE_LIST, READER, CHANNEL, RAW]",
         var_type);
   }
 }

paddle/fluid/framework/framework.proto

@@ -113,7 +113,10 @@ message VarType {
     PLACE_LIST = 14;
     READER = 15;
     CHANNEL = 16;
-    NCCL_COM = 17;
+    // Any runtime decided variable type is raw
+    // raw variables should manage their own allocations
+    // in operators like nccl_op
+    RAW = 17;
   }
 
   required Type type = 1;

paddle/fluid/operators/detail/grpc_client.cc

@@ -177,8 +177,8 @@ std::shared_ptr<grpc::Channel> RPCClient::GetChannel(const std::string& ep) {
   args.SetMaxSendMessageSize(std::numeric_limits<int>::max());
   args.SetMaxReceiveMessageSize(std::numeric_limits<int>::max());
 
-  auto ch = std::shared_ptr<grpc::Channel>(
-      grpc::CreateCustomChannel(ep, grpc::InsecureChannelCredentials(), args));
+  auto ch =
+      grpc::CreateCustomChannel(ep, grpc::InsecureChannelCredentials(), args);
 
   channels_[ep] = ch;
   return ch;

paddle/fluid/operators/nccl_op.cc

@@ -65,7 +65,7 @@ class NCCLInitOpVarTypeInference : public framework::VarTypeInference {
                   framework::BlockDesc *block) const override {
     auto out_var_name = op_desc.Output("Communicator").front();
     auto &out_var = block->FindRecursiveOrCreateVar(out_var_name);
-    auto var_type = framework::proto::VarType::NCCL_COM;
+    auto var_type = framework::proto::VarType::RAW;
     out_var.SetType(var_type);
   }
 };

paddle/fluid/operators/send_op.cc

@@ -121,9 +121,27 @@ This operator will send tensor to recv_op at the parameter server.
   }
 };
 
+class SendOpVarTypeInference : public framework::VarTypeInference {
+ public:
+  void operator()(const framework::OpDesc& op_desc,
+                  framework::BlockDesc* block) const override {
+    auto out_var_name = op_desc.Output("RPCClient").front();
+    auto& out_var = block->FindRecursiveOrCreateVar(out_var_name);
+    auto var_type = framework::proto::VarType::RAW;
+    out_var.SetType(var_type);
+  }
+};
+
+class SendOpShapeInference : public framework::InferShapeBase {
+ public:
+  void operator()(framework::InferShapeContext* ctx) const override {}
+};
+
 }  // namespace operators
 }  // namespace paddle
 
 namespace ops = paddle::operators;
 
-REGISTER_OPERATOR(send, ops::SendOp, ops::SendOpMaker);
+REGISTER_OPERATOR(send, ops::SendOp, paddle::framework::EmptyGradOpMaker,
+                  ops::SendOpMaker, ops::SendOpVarTypeInference,
+                  ops::SendOpShapeInference);

paddle/fluid/pybind/protobuf.cc

@@ -252,7 +252,7 @@ void BindVarDsec(py::module &m) {
       .value("CHANNEL", proto::VarType::CHANNEL)
       .value("PLACE_LIST", proto::VarType::PLACE_LIST)
       .value("READER", proto::VarType::READER)
-      .value("NCCL_COM", proto::VarType::NCCL_COM);
+      .value("RAW", proto::VarType::RAW);
 }
 
 void BindOpDesc(py::module &m) {

paddle/testing/paddle_gtest_main.cc

@@ -28,10 +28,9 @@ int main(int argc, char** argv) {
   }
 #ifdef PADDLE_WITH_CUDA
   new_argv.push_back(
-      strdup("--tryfromenv=fraction_of_gpu_memory_to_use,use_pinned_memory,"
-             "warpctc_dir"));
+      strdup("--tryfromenv=fraction_of_gpu_memory_to_use,use_pinned_memory"));
 #else
-  new_argv.push_back(strdup("--tryfromenv=use_pinned_memory,warpctc_dir"));
+  new_argv.push_back(strdup("--tryfromenv=use_pinned_memory"));
 #endif
   int new_argc = static_cast<int>(new_argv.size());
   char** new_argv_address = new_argv.data();

python/paddle/fluid/.gitignore

 proto
+core.so

python/paddle/fluid/distribute_transpiler.py

@@ -226,8 +226,7 @@ class DistributeTranspiler:
         rpc_client_var = program.global_block().create_var(
             name="RPC_CLIENT_VAR",
             persistable=True,
-            dtype='float32',  # dtype and shape is not used in fact
-            shape=[0])
+            type=core.VarDesc.VarType.RAW)
 
         # create send_op
         program.global_block().append_op(

python/paddle/fluid/layers/layer_function_generator.py

@@ -130,7 +130,7 @@ def generate_layer_fn(op_type):
     o_name = not_intermediate_outputs[0].name
     intermediate_output_names = [output.name for output in intermediate_outputs]
 
-    def infer_and_check_dtype(op_proto, **kwargs):
+    def infer_and_check_dtype(op_proto, *args, **kwargs):
         """
         This function performs the sanity check for dtype and
         instance type.
@@ -141,6 +141,10 @@ def generate_layer_fn(op_type):
             val = kwargs.pop(name, [])
             if not isinstance(val, list) and not isinstance(val, tuple):
                 val = [val]
+            if len(val) == 0:
+                val = [args[0]]
+                args = args[1:]
+
             for each in val:
                 if not isinstance(each, Variable):
                     raise ValueError("input of {0} must be variable".format(
@@ -155,10 +159,10 @@ def generate_layer_fn(op_type):
 
         return dtype
 
-    def func(**kwargs):
+    def func(*args, **kwargs):
         helper = LayerHelper(op_type, **kwargs)
 
-        dtype = infer_and_check_dtype(op_proto, **kwargs)
+        dtype = infer_and_check_dtype(op_proto, *args, **kwargs)
 
         inputs = dict()
         for ipt in op_proto.inputs:
@@ -166,6 +170,9 @@ def generate_layer_fn(op_type):
             val = kwargs.pop(name, [])
             if not isinstance(val, list) and not isinstance(val, tuple):
                 val = [val]
+            if len(val) == 0 and len(args) != 0:
+                val = args[0]
+                args = args[1:]
             inputs[ipt.name] = val
 
         outputs = dict()

python/paddle/fluid/layers/tensor.py

@@ -160,8 +160,8 @@ def sums(input, out=None):
           a0 = layers.array_read(array=tmp, i=i)
           i = layers.increment(x=i)
           a1 = layers.array_read(array=tmp, i=i)
-          mean_a0 = layers.mean(x=a0)
-          mean_a1 = layers.mean(x=a1)
+          mean_a0 = layers.mean(a0)
+          mean_a1 = layers.mean(a1)
           a_sum = layers.sums(input=[mean_a0, mean_a1])
     """
     helper = LayerHelper('sum', **locals())

python/paddle/fluid/optimizer.py

@@ -36,10 +36,18 @@ class Optimizer(object):
     """
 
     def __init__(self, learning_rate, global_step=None, regularization=None):
-        assert learning_rate is not None
+        if not isinstance(learning_rate, float) and \
+                not isinstance(learning_rate, framework.Variable):
+            raise TypeError("learning rate should be float or Variable")
         self._global_step = global_step
         self.regularization = regularization
-        self._global_learning_rate = learning_rate
+        self._learning_rate = learning_rate
+        # each program should have a independent learning rate
+        # program -> Variable(learning_rate)
+        self._learning_rate_map = dict()
+        if isinstance(self._learning_rate, framework.Variable):
+            self._learning_rate_map[framework.default_main_program(
+            )] = self._learning_rate
         # Dictionary of accumulators. Some optimizer subclasses need to
         # allocate and manage extra variables associated with the parameters
         # to train. These variables are called accumulators.
@@ -48,26 +56,33 @@ class Optimizer(object):
         self.helper = None
 
     def _create_global_learning_rate(self):
-        if isinstance(self._global_learning_rate, float):
-            self._global_learning_rate = layers.create_global_var(
+        lr = self.global_learning_rate()
+
+        if isinstance(lr, framework.Variable):
+            return
+        else:
+            if not isinstance(self._learning_rate, float):
+                raise TypeError(
+                    "learning rate variable is create outside optimizer,"
+                    "can not create new learning rate variable for new program")
+
+        # create learning rate in the current main program
+        self._learning_rate_map[framework.default_main_program(
+        )] = layers.create_global_var(
             name=unique_name.generate("learning_rate"),
             shape=[1],
-                value=float(self._global_learning_rate),
+            value=float(self._learning_rate),
             dtype='float32',
             persistable=True)
 
-        if not isinstance(self._global_learning_rate, framework.Variable):
-            raise ValueError("learning rate should be a Variable, "
-                             "actual type is %s",
-                             type(self._global_learning_rate))
-
-    @property
-    def global_learning_rate(self):
+    def global_learning_rate(self, program=None):
         """
         get global decayed learning rate
         :return:
         """
-        return self._global_learning_rate
+        if program is None:
+            program = framework.default_main_program()
+        return self._learning_rate_map.get(program, None)
 
     def _append_optimize_op(self, block, param_and_grad):
         """ append optimize operator to block and return all the added optimize_op
@@ -78,7 +93,7 @@ class Optimizer(object):
         # create learning rate variable for every parameter
         param = param_and_grad[0]
         param_lr = param.optimize_attr['learning_rate']
-        return self._global_learning_rate * param_lr
+        return self.global_learning_rate() * param_lr
 
     def _create_accumulators(self, block, parameters):
         """Create all accumulators needed by the parameters

python/paddle/fluid/tests/book/notest_rnn_encoder_decoer.py

@@ -147,7 +147,7 @@ def seq_to_seq_net():
     label = fluid.layers.data(
         name='label_sequence', shape=[1], dtype='int64', lod_level=1)
     cost = fluid.layers.cross_entropy(input=prediction, label=label)
-    avg_cost = fluid.layers.mean(x=cost)
+    avg_cost = fluid.layers.mean(cost)
 
     return avg_cost, prediction
 

python/paddle/fluid/tests/book/test_fit_a_line.py

@@ -29,7 +29,7 @@ def train(use_cuda, save_dirname):
     y = fluid.layers.data(name='y', shape=[1], dtype='float32')
 
     cost = fluid.layers.square_error_cost(input=y_predict, label=y)
-    avg_cost = fluid.layers.mean(x=cost)
+    avg_cost = fluid.layers.mean(cost)
 
     sgd_optimizer = fluid.optimizer.SGD(learning_rate=0.001)
     sgd_optimizer.minimize(avg_cost)

python/paddle/fluid/tests/book/test_image_classification.py

@@ -110,7 +110,7 @@ def train(net_type, use_cuda, save_dirname):
 
     predict = fluid.layers.fc(input=net, size=classdim, act='softmax')
     cost = fluid.layers.cross_entropy(input=predict, label=label)
-    avg_cost = fluid.layers.mean(x=cost)
+    avg_cost = fluid.layers.mean(cost)
     acc = fluid.layers.accuracy(input=predict, label=label)
 
     # Test program 

python/paddle/fluid/tests/book/test_label_semantic_roles.py

@@ -164,7 +164,7 @@ def train(use_cuda, save_dirname=None):
         label=target,
         param_attr=fluid.ParamAttr(
             name='crfw', learning_rate=mix_hidden_lr))
-    avg_cost = fluid.layers.mean(x=crf_cost)
+    avg_cost = fluid.layers.mean(crf_cost)
 
     # TODO(qiao)
     # check other optimizers and check why out will be NAN

python/paddle/fluid/tests/book/test_machine_translation.py

@@ -178,7 +178,7 @@ def train_main(use_cuda, is_sparse):
     label = pd.data(
         name="target_language_next_word", shape=[1], dtype='int64', lod_level=1)
     cost = pd.cross_entropy(input=rnn_out, label=label)
-    avg_cost = pd.mean(x=cost)
+    avg_cost = pd.mean(cost)
 
     optimizer = fluid.optimizer.Adagrad(learning_rate=1e-4)
     optimizer.minimize(avg_cost)

python/paddle/fluid/tests/book/test_recognize_digits.py

@@ -48,7 +48,7 @@ BATCH_SIZE = 64
 def loss_net(hidden, label):
     prediction = fluid.layers.fc(input=hidden, size=10, act='softmax')
     loss = fluid.layers.cross_entropy(input=prediction, label=label)
-    avg_loss = fluid.layers.mean(x=loss)
+    avg_loss = fluid.layers.mean(loss)
     acc = fluid.layers.accuracy(input=prediction, label=label)
     return prediction, avg_loss, acc
 
@@ -101,8 +101,8 @@ def train(nn_type, use_cuda, parallel, save_dirname, save_param_filename):
 
         avg_loss, acc = pd()
         # get mean loss and acc through every devices.
-        avg_loss = fluid.layers.mean(x=avg_loss)
-        acc = fluid.layers.mean(x=acc)
+        avg_loss = fluid.layers.mean(avg_loss)
+        acc = fluid.layers.mean(acc)
     else:
         prediction, avg_loss, acc = net_conf(img, label)
 

python/paddle/fluid/tests/book/test_recommender_system.py

@@ -147,7 +147,7 @@ def model():
 
     label = layers.data(name='score', shape=[1], dtype='float32')
     square_cost = layers.square_error_cost(input=scale_infer, label=label)
-    avg_cost = layers.mean(x=square_cost)
+    avg_cost = layers.mean(square_cost)
 
     return scale_infer, avg_cost
 

python/paddle/fluid/tests/book/test_understand_sentiment.py

@@ -42,7 +42,7 @@ def convolution_net(data, label, input_dim, class_dim=2, emb_dim=32,
                                  size=class_dim,
                                  act="softmax")
     cost = fluid.layers.cross_entropy(input=prediction, label=label)
-    avg_cost = fluid.layers.mean(x=cost)
+    avg_cost = fluid.layers.mean(cost)
     accuracy = fluid.layers.accuracy(input=prediction, label=label)
     return avg_cost, accuracy, prediction
 
@@ -82,7 +82,7 @@ def dyn_rnn_lstm(data, label, input_dim, class_dim=2, emb_dim=32,
     last = fluid.layers.sequence_last_step(rnn())
     prediction = fluid.layers.fc(input=last, size=class_dim, act="softmax")
     cost = fluid.layers.cross_entropy(input=prediction, label=label)
-    avg_cost = fluid.layers.mean(x=cost)
+    avg_cost = fluid.layers.mean(cost)
     accuracy = fluid.layers.accuracy(input=prediction, label=label)
     return avg_cost, accuracy, prediction
 
@@ -119,7 +119,7 @@ def stacked_lstm_net(data,
                                  size=class_dim,
                                  act='softmax')
     cost = fluid.layers.cross_entropy(input=prediction, label=label)
-    avg_cost = fluid.layers.mean(x=cost)
+    avg_cost = fluid.layers.mean(cost)
     accuracy = fluid.layers.accuracy(input=prediction, label=label)
     return avg_cost, accuracy, prediction
 
@@ -158,8 +158,8 @@ def train(word_dict, net_method, use_cuda, parallel=False, save_dirname=None):
             pd.write_output(acc)
 
         cost, acc = pd()
-        cost = fluid.layers.mean(x=cost)
-        acc_out = fluid.layers.mean(x=acc)
+        cost = fluid.layers.mean(cost)
+        acc_out = fluid.layers.mean(acc)
         prediction = None
         assert save_dirname is None
 

python/paddle/fluid/tests/book/test_word2vec.py

@@ -118,7 +118,7 @@ def train(use_cuda, is_sparse, parallel, save_dirname):
                                        size=dict_size,
                                        act='softmax')
         cost = fluid.layers.cross_entropy(input=predict_word, label=words[4])
-        avg_cost = fluid.layers.mean(x=cost)
+        avg_cost = fluid.layers.mean(cost)
         return avg_cost, predict_word
 
     word_dict = paddle.dataset.imikolov.build_dict()
@@ -143,7 +143,7 @@ def train(use_cuda, is_sparse, parallel, save_dirname):
                 ]))
             pd.write_output(avg_cost)
 
-        avg_cost = fluid.layers.mean(x=pd())
+        avg_cost = fluid.layers.mean(pd())
 
     sgd_optimizer = fluid.optimizer.SGD(learning_rate=0.001)
     sgd_optimizer.minimize(avg_cost)

python/paddle/fluid/tests/book_distribute/notest_dist_fit_a_line.py

@@ -24,7 +24,7 @@ y_predict = fluid.layers.fc(input=x, size=1, act=None)
 y = fluid.layers.data(name='y', shape=[1], dtype='float32')
 
 cost = fluid.layers.square_error_cost(input=y_predict, label=y)
-avg_cost = fluid.layers.mean(x=cost)
+avg_cost = fluid.layers.mean(cost)
 
 sgd_optimizer = fluid.optimizer.SGD(learning_rate=0.001)
 optimize_ops, params_grads = sgd_optimizer.minimize(avg_cost)

python/paddle/fluid/tests/book_distribute/notest_dist_image_classification.py

@@ -114,7 +114,7 @@ else:
 
 predict = fluid.layers.fc(input=net, size=classdim, act='softmax')
 cost = fluid.layers.cross_entropy(input=predict, label=label)
-avg_cost = fluid.layers.mean(x=cost)
+avg_cost = fluid.layers.mean(cost)
 
 optimizer = fluid.optimizer.Adam(learning_rate=0.001)
 optimize_ops, params_grads = optimizer.minimize(avg_cost)

python/paddle/fluid/tests/book_distribute/notest_dist_label_semantic_roles.py

@@ -154,7 +154,7 @@ def main():
         label=target,
         param_attr=fluid.ParamAttr(
             name='crfw', learning_rate=mix_hidden_lr))
-    avg_cost = fluid.layers.mean(x=crf_cost)
+    avg_cost = fluid.layers.mean(crf_cost)
 
     # TODO(qiao)
     # check other optimizers and check why out will be NAN

python/paddle/fluid/tests/book_distribute/notest_dist_word2vec.py

@@ -65,7 +65,7 @@ concat_embed = fluid.layers.concat(
 hidden1 = fluid.layers.fc(input=concat_embed, size=HIDDEN_SIZE, act='sigmoid')
 predict_word = fluid.layers.fc(input=hidden1, size=dict_size, act='softmax')
 cost = fluid.layers.cross_entropy(input=predict_word, label=next_word)
-avg_cost = fluid.layers.mean(x=cost)
+avg_cost = fluid.layers.mean(cost)
 sgd_optimizer = fluid.optimizer.SGD(learning_rate=0.001)
 optimize_ops, params_grads = sgd_optimizer.minimize(avg_cost)
 train_reader = paddle.batch(

python/paddle/fluid/tests/book_distribute/notest_machine_translation.py

@@ -94,7 +94,7 @@ def main():
     label = layers.data(
         name="target_language_next_word", shape=[1], dtype='int64', lod_level=1)
     cost = layers.cross_entropy(input=rnn_out, label=label)
-    avg_cost = fluid.layers.mean(x=cost)
+    avg_cost = fluid.layers.mean(cost)
 
     optimizer = fluid.optimizer.Adagrad(learning_rate=1e-4)
     optimize_ops, params_grads = optimizer.minimize(avg_cost)

python/paddle/fluid/tests/book_distribute/notest_recognize_digits_conv_dist.py

@@ -37,7 +37,7 @@ conv_pool_2 = fluid.nets.simple_img_conv_pool(
 
 predict = fluid.layers.fc(input=conv_pool_2, size=10, act="softmax")
 cost = fluid.layers.cross_entropy(input=predict, label=label)
-avg_cost = fluid.layers.mean(x=cost)
+avg_cost = fluid.layers.mean(cost)
 optimizer = fluid.optimizer.Adam(learning_rate=0.01)
 optimize_ops, params_grads = optimizer.minimize(avg_cost)
 

python/paddle/fluid/tests/book_distribute/notest_recognize_digits_mlp_dist.py

@@ -32,7 +32,7 @@ predict = fluid.layers.fc(input=hidden2, size=10, act='softmax')
 label = fluid.layers.data(name='y', shape=[1], dtype='int64')
 
 cost = fluid.layers.cross_entropy(input=predict, label=label)
-avg_cost = fluid.layers.mean(x=cost)
+avg_cost = fluid.layers.mean(cost)
 
 optimizer = fluid.optimizer.Momentum(learning_rate=0.001, momentum=0.9)
 optimize_ops, params_grads = optimizer.minimize(avg_cost)

python/paddle/fluid/tests/book_distribute/notest_recommender_system_dist.py

@@ -117,7 +117,7 @@ def model():
 
     label = layers.data(name='score', shape=[1], dtype='float32')
     square_cost = layers.square_error_cost(input=scale_infer, label=label)
-    avg_cost = layers.mean(x=square_cost)
+    avg_cost = layers.mean(square_cost)
 
     return avg_cost
 

python/paddle/fluid/tests/book_distribute/notest_understand_sentiment_conv_dist.py

@@ -38,7 +38,7 @@ def convolution_net(data, label, input_dim, class_dim=2, emb_dim=32,
                                  size=class_dim,
                                  act="softmax")
     cost = fluid.layers.cross_entropy(input=prediction, label=label)
-    avg_cost = fluid.layers.mean(x=cost)
+    avg_cost = fluid.layers.mean(cost)
     adam_optimizer = fluid.optimizer.Adam(learning_rate=0.002)
     optimize_ops, params_grads = adam_optimizer.minimize(avg_cost)
     accuracy = fluid.evaluator.Accuracy(input=prediction, label=label)

python/paddle/fluid/tests/book_distribute/notest_understand_sentiment_dynamic_lstm.py

@@ -49,7 +49,7 @@ def stacked_lstm_net(data,
                                  size=class_dim,
                                  act='softmax')
     cost = fluid.layers.cross_entropy(input=prediction, label=label)
-    avg_cost = fluid.layers.mean(x=cost)
+    avg_cost = fluid.layers.mean(cost)
     adam_optimizer = fluid.optimizer.Adam(learning_rate=0.002)
     optimize_ops, params_grads = adam_optimizer.minimize(avg_cost)
     accuracy = fluid.evaluator.Accuracy(input=prediction, label=label)

python/paddle/fluid/tests/book_memory_optimization/test_memopt_image_classification_train.py

@@ -117,7 +117,7 @@ else:
 
 predict = fluid.layers.fc(input=net, size=classdim, act='softmax')
 cost = fluid.layers.cross_entropy(input=predict, label=label)
-avg_cost = fluid.layers.mean(x=cost)
+avg_cost = fluid.layers.mean(cost)
 
 optimizer = fluid.optimizer.Adam(learning_rate=0.001)
 opts = optimizer.minimize(avg_cost)

python/paddle/fluid/tests/book_memory_optimization/test_memopt_machine_translation.py

@@ -100,7 +100,7 @@ def main():
     label = layers.data(
         name="target_language_next_word", shape=[1], dtype='int64', lod_level=1)
     cost = layers.cross_entropy(input=rnn_out, label=label)
-    avg_cost = fluid.layers.mean(x=cost)
+    avg_cost = fluid.layers.mean(cost)
 
     optimizer = fluid.optimizer.Adagrad(learning_rate=1e-4)
     optimizer.minimize(avg_cost)

python/paddle/fluid/tests/demo/fc_gan.py

@@ -96,7 +96,7 @@ def main():
             x=D(img),
             label=fluid.layers.data(
                 name='label', shape=[1], dtype='float32'))
-        d_loss = fluid.layers.mean(x=d_loss)
+        d_loss = fluid.layers.mean(d_loss)
 
     with fluid.program_guard(dg_program, startup_program):
         noise = fluid.layers.data(
@@ -107,7 +107,7 @@ def main():
             x=D(g_img),
             label=fluid.layers.fill_constant_batch_size_like(
                 input=noise, dtype='float32', shape=[-1, 1], value=1.0))
-        dg_loss = fluid.layers.mean(x=dg_loss)
+        dg_loss = fluid.layers.mean(dg_loss)
 
     opt = fluid.optimizer.Adam(learning_rate=LEARNING_RATE)
 

python/paddle/fluid/tests/test_error_clip.py

@@ -33,7 +33,7 @@ with fluid.program_guard(main_program=prog):
     label = fluid.layers.data(name='y', shape=[1], dtype='int64')
 
     cost = fluid.layers.cross_entropy(input=predict, label=label)
-    avg_cost = fluid.layers.mean(x=cost)
+    avg_cost = fluid.layers.mean(cost)
 
 prog_clip = prog.clone()
 prog_clip.block(0).var(hidden1.name).set_error_clip(

python/paddle/fluid/tests/test_gradient_clip.py

@@ -30,7 +30,7 @@ with fluid.program_guard(main_program=prog):
     label = fluid.layers.data(name='y', shape=[1], dtype='int64')
 
     cost = fluid.layers.cross_entropy(input=predict, label=label)
-    avg_cost = fluid.layers.mean(x=cost)
+    avg_cost = fluid.layers.mean(cost)
 
 prog_clip = prog.clone()
 

python/paddle/fluid/tests/test_mnist_if_else_op.py

@@ -56,7 +56,7 @@ class TestMNISTIfElseOp(unittest.TestCase):
             prob = layers.merge_lod_tensor(
                 in_true=true_out, in_false=false_out, mask=cond, x=image)
             loss = layers.cross_entropy(input=prob, label=label)
-            avg_loss = layers.mean(x=loss)
+            avg_loss = layers.mean(loss)
 
             optimizer = MomentumOptimizer(learning_rate=0.001, momentum=0.9)
             optimizer.minimize(avg_loss, startup_prog)
@@ -113,7 +113,7 @@ class TestMNISTIfElseOp(unittest.TestCase):
 
             prob = ie()
             loss = layers.cross_entropy(input=prob[0], label=label)
-            avg_loss = layers.mean(x=loss)
+            avg_loss = layers.mean(loss)
 
             optimizer = MomentumOptimizer(learning_rate=0.001, momentum=0.9)
             optimizer.minimize(avg_loss, startup_prog)

python/paddle/fluid/tests/unittests/test_array_read_write_op.py

@@ -49,15 +49,15 @@ class TestArrayReadWrite(unittest.TestCase):
         i = layers.increment(x=i)
         a2 = layers.array_read(array=arr, i=i)
 
-        mean_a0 = layers.mean(x=a0)
-        mean_a1 = layers.mean(x=a1)
-        mean_a2 = layers.mean(x=a2)
+        mean_a0 = layers.mean(a0)
+        mean_a1 = layers.mean(a1)
+        mean_a2 = layers.mean(a2)
 
         a_sum = layers.sums(input=[mean_a0, mean_a1, mean_a2])
 
-        mean_x0 = layers.mean(x=x[0])
-        mean_x1 = layers.mean(x=x[1])
-        mean_x2 = layers.mean(x=x[2])
+        mean_x0 = layers.mean(x[0])
+        mean_x1 = layers.mean(x[1])
+        mean_x2 = layers.mean(x[2])
 
         x_sum = layers.sums(input=[mean_x0, mean_x1, mean_x2])
 

python/paddle/fluid/tests/unittests/test_calc_gradient.py

@@ -26,7 +26,7 @@ class TestCalcGradient(unittest.TestCase):
         x = layers.create_parameter(dtype="float32", shape=[5, 10])
         y = layers.create_parameter(dtype="float32", shape=[10, 8])
         mul_out = layers.mul(x=x, y=y)
-        mean_out = layers.mean(x=mul_out)
+        mean_out = layers.mean(mul_out)
         a = calc_gradient(mean_out, mul_out)
         b = calc_gradient(mean_out, x)
         place = fluid.CPUPlace()

python/paddle/fluid/tests/unittests/test_conditional_block.py

@@ -39,7 +39,7 @@ class ConditionalBlock(unittest.TestCase):
 
         outs = exe.run(feed={'X': x}, fetch_list=[out])[0]
         print outs
-        loss = layers.mean(x=out)
+        loss = layers.mean(out)
         append_backward(loss=loss)
         outs = exe.run(
             feed={'X': x},

python/paddle/fluid/tests/unittests/test_dyn_rnn.py

@@ -81,7 +81,7 @@ class TestDynRNN(unittest.TestCase):
             logits = fluid.layers.fc(input=last, size=1, act=None)
             loss = fluid.layers.sigmoid_cross_entropy_with_logits(
                 x=logits, label=label)
-            loss = fluid.layers.mean(x=loss)
+            loss = fluid.layers.mean(loss)
             sgd = fluid.optimizer.SGD(1e-4)
             sgd.minimize(loss=loss)
         cpu = fluid.CPUPlace()
@@ -119,7 +119,7 @@ class TestDynRNN(unittest.TestCase):
             label = fluid.layers.data(name='label', shape=[1], dtype='float32')
             loss = fluid.layers.sigmoid_cross_entropy_with_logits(
                 x=logits, label=label)
-            loss = fluid.layers.mean(x=loss)
+            loss = fluid.layers.mean(loss)
             sgd = fluid.optimizer.Adam(1e-3)
             sgd.minimize(loss=loss)
 

python/paddle/fluid/tests/unittests/test_dynrnn_gradient_check.py

@@ -272,7 +272,7 @@ class TestSimpleMul(SeedFixedTestCase):
 
         out = rnn()
         out = fluid.layers.sequence_pool(out, pool_type='last')
-        loss = fluid.layers.mean(x=out)
+        loss = fluid.layers.mean(out)
         fluid.backward.append_backward(loss)
 
         cpu = fluid.CPUPlace()
@@ -348,7 +348,7 @@ class TestSimpleMulWithMemory(SeedFixedTestCase):
 
         out = rnn()
         last = fluid.layers.sequence_pool(input=out, pool_type='last')
-        loss = fluid.layers.mean(x=last)
+        loss = fluid.layers.mean(last)
         fluid.backward.append_backward(loss)
 
         cpu = fluid.CPUPlace()

python/paddle/fluid/tests/unittests/test_dynrnn_static_input.py

@@ -125,7 +125,7 @@ class TestDyRnnStaticInput(unittest.TestCase):
             return static_input_step_outs
 
         last = fluid.layers.sequence_pool(input=rnn(), pool_type='last')
-        loss = fluid.layers.mean(x=last)
+        loss = fluid.layers.mean(last)
         append_backward(loss)
         static_input_grad = self._program.global_block().var(
             framework.grad_var_name('static_input_tensor'))

python/paddle/fluid/tests/unittests/test_inference_model_io.py

@@ -38,7 +38,7 @@ class TestBook(unittest.TestCase):
             y_predict = layers.fc(input=x, size=1, act=None)
 
             cost = layers.square_error_cost(input=y_predict, label=y)
-            avg_cost = layers.mean(x=cost)
+            avg_cost = layers.mean(cost)
 
             sgd_optimizer = optimizer.SGDOptimizer(learning_rate=0.001)
             sgd_optimizer.minimize(avg_cost, init_program)

python/paddle/fluid/tests/unittests/test_layers.py

@@ -30,7 +30,7 @@ class TestBook(unittest.TestCase):
             y_predict = layers.fc(input=x, size=1, act=None)
             y = layers.data(name='y', shape=[1], dtype='float32')
             cost = layers.square_error_cost(input=y_predict, label=y)
-            avg_cost = layers.mean(x=cost)
+            avg_cost = layers.mean(cost)
             self.assertIsNotNone(avg_cost)
             program.append_backward(avg_cost)
 
@@ -49,7 +49,7 @@ class TestBook(unittest.TestCase):
                                 act='softmax',
                                 param_attr=["sftmax.w1", "sftmax.w2"])
             cost = layers.cross_entropy(input=predict, label=label)
-            avg_cost = layers.mean(x=cost)
+            avg_cost = layers.mean(cost)
             self.assertIsNotNone(avg_cost)
 
         print(str(program))
@@ -92,7 +92,7 @@ class TestBook(unittest.TestCase):
 
             predict = layers.fc(input=conv_pool_2, size=10, act="softmax")
             cost = layers.cross_entropy(input=predict, label=label)
-            avg_cost = layers.mean(x=cost)
+            avg_cost = layers.mean(cost)
 
             program.append_backward(avg_cost)
 
@@ -140,7 +140,7 @@ class TestBook(unittest.TestCase):
                                      size=dict_size,
                                      act='softmax')
             cost = layers.cross_entropy(input=predict_word, label=next_word)
-            avg_cost = layers.mean(x=cost)
+            avg_cost = layers.mean(cost)
             self.assertIsNotNone(avg_cost)
 
         print(str(program))
@@ -287,7 +287,7 @@ class TestBook(unittest.TestCase):
                           num_total_classes=dict_size,
                           param_attr='nce.w',
                           bias_attr='nce.b')
-        avg_loss = layers.mean(x=loss)
+        avg_loss = layers.mean(loss)
         self.assertIsNotNone(avg_loss)
         print(str(default_main_program()))
 

python/paddle/fluid/tests/unittests/test_lod_tensor_array_ops.py

@@ -182,7 +182,7 @@ class TestCPULoDTensorArrayOpGrad(unittest.TestCase):
             array = layers.lod_tensor_to_array(x, table)
             result = layers.array_to_lod_tensor(array, table)
 
-            mean = layers.mean(x=result)
+            mean = layers.mean(result)
 
             append_backward(mean)
 

python/paddle/fluid/tests/unittests/test_memory_optimization_transpiler.py

@@ -29,7 +29,7 @@ class TestControlFlowGraph(unittest.TestCase):
             y_predict = layers.fc(input=x, size=1, act=None)
             y = layers.data(name='y', shape=[1], dtype='float32')
             cost = layers.square_error_cost(input=y_predict, label=y)
-            avg_cost = layers.mean(x=cost)
+            avg_cost = layers.mean(cost)
             opt = optimizer.SGD(learning_rate=0.001)
             opt = opt.minimize(avg_cost)
 

python/paddle/fluid/tests/unittests/test_parallel_op.py

@@ -127,7 +127,7 @@ class BaseParallelForTest(unittest.TestCase):
                 data = next(generator)
                 loss = generator.send(data)
             self.assertIsNotNone(loss)
-            avg_loss = fluid.layers.mean(x=loss)
+            avg_loss = fluid.layers.mean(loss)
             fluid.backward.append_backward(loss=avg_loss)
 
         exe = fluid.Executor(place)
@@ -170,7 +170,7 @@ class ParallelOpTest(BaseParallelForTest):
         x = fluid.layers.data(shape=[784], dtype='float32', name='img')
         x = yield x
         hidden = fluid.layers.fc(input=x, size=200, param_attr='fc1.w')
-        loss = fluid.layers.mean(x=hidden)
+        loss = fluid.layers.mean(hidden)
         yield loss
 
     def test_simple_fc(self):
@@ -200,7 +200,7 @@ class ParallelOpTestMultipleInput(BaseParallelForTest):
         hidden1 = fluid.layers.fc(input=x, size=200, param_attr='fc1.w')
         hidden2 = fluid.layers.fc(input=hidden1, size=200, param_attr='fc2.w')
         hidden3 = fluid.layers.fc(input=hidden2, size=200, param_attr='fc3.w')
-        loss = fluid.layers.mean(x=hidden3)
+        loss = fluid.layers.mean(hidden3)
         yield loss
 
     def test_simple_fc(self):

python/paddle/fluid/tests/unittests/test_print_op.py

@@ -35,7 +35,7 @@ class TestPrintOpCPU(unittest.TestCase):
         x.stop_gradient = False
         printed = layers.Print(input=x, **kargs)
         if only_forward: return printed
-        loss = layers.mean(x=printed)
+        loss = layers.mean(printed)
         append_backward(loss=loss)
         return loss
 

python/paddle/fluid/tests/unittests/test_profiler.py

@@ -54,7 +54,7 @@ class TestProfiler(unittest.TestCase):
             predict = fluid.layers.fc(input=hidden2, size=10, act='softmax')
             label = fluid.layers.data(name='y', shape=[1], dtype='int64')
             cost = fluid.layers.cross_entropy(input=predict, label=label)
-            avg_cost = fluid.layers.mean(x=cost)
+            avg_cost = fluid.layers.mean(cost)
             accuracy = fluid.evaluator.Accuracy(input=predict, label=label)
 
         optimizer = fluid.optimizer.Momentum(learning_rate=0.001, momentum=0.9)

python/paddle/fluid/tests/unittests/test_recurrent_op.py

@@ -127,7 +127,7 @@ class RecurrentOpTest1(unittest.TestCase):
         self.output_shape = (self.sent_len, self.batch_size, self.input_dim)
         self.py_rnn = PySimpleRNN1(self.input_shape, self.output_shape)
 
-        self.output = layers.mean(x=self.create_rnn_op(), **self.p_info)
+        self.output = layers.mean(self.create_rnn_op(), **self.p_info)
 
     def create_rnn_op(self):
         x = layers.data(
@@ -261,7 +261,7 @@ class RecurrentOpTest2(RecurrentOpTest1):
         self.output_shape = (self.sent_len, self.batch_size, self.input_dim)
         self.py_rnn = PySimpleRNN2(self.input_shape, self.output_shape)
 
-        self.output = layers.mean(x=self.create_rnn_op(), **self.p_info)
+        self.output = layers.mean(self.create_rnn_op(), **self.p_info)
 
     def create_rnn_op(self):
         x = layers.data(
@@ -360,7 +360,7 @@ class RecurrentOpMultipleMemoryTest(RecurrentOpTest1):
         self.py_rnn = RecurrentOpMultipleMemoryTest.PySimpleRNN3(
             self.input_shape, self.output_shape)
 
-        self.output = layers.mean(x=self.create_rnn_op(), **self.p_info)
+        self.output = layers.mean(self.create_rnn_op(), **self.p_info)
 
     def create_rnn_op(self):
         x = layers.data(
@@ -444,7 +444,7 @@ class RecurrentOpNoMemBootTest(RecurrentOpTest1):
         self.output_shape = (self.sent_len, self.batch_size, self.input_dim)
         self.py_rnn = RecurrentOpNoMemBootTest.PySimpleRNN4(self.input_shape,
                                                             self.output_shape)
-        self.output = layers.mean(x=self.create_rnn_op(), **self.p_info)
+        self.output = layers.mean(self.create_rnn_op(), **self.p_info)
         print self.main_program
 
     def create_rnn_op(self):

python/paddle/fluid/tests/unittests/test_registry.py

@@ -22,7 +22,7 @@ class TestRegistry(unittest.TestCase):
     @decorators.prog_scope()
     def test_registry_layer(self):
         x = fluid.layers.data(name='X', shape=[10, 10], dtype='float32')
-        output = fluid.layers.mean(x=x)
+        output = fluid.layers.mean(x)
 
         place = fluid.CPUPlace()
         exe = fluid.Executor(place)

python/paddle/fluid/tests/unittests/test_shrink_rnn_memory.py

@@ -39,7 +39,7 @@ class TestShrinkRNNMemoryBase(unittest.TestCase):
         i = layers.increment(x=i)
         i.stop_gradient = True
         self.mem3 = layers.shrink_memory(x=self.mem2, i=i, table=table)
-        mem3_mean = layers.mean(x=self.mem3)
+        mem3_mean = layers.mean(self.mem3)
         append_backward(loss=mem3_mean)
         self.x_grad = self.main_program.global_block().var('x@GRAD')
 

python/paddle/fluid/tests/unittests/test_split_and_merge_lod_tensor_op.py

@@ -145,7 +145,7 @@ class TestCPUSplitMergeLoDTensorGrad(unittest.TestCase):
                 input=x, mask=y, level=level)
             out = layers.merge_lod_tensor(
                 in_true=out_true, in_false=out_false, mask=y, x=x, level=level)
-            mean = layers.mean(x=out)
+            mean = layers.mean(out)
 
             append_backward(mean)
 

python/paddle/fluid/tests/unittests/test_while_op.py

@@ -58,7 +58,7 @@ class TestWhileOp(unittest.TestCase):
             layers.less_than(x=i, y=array_len, cond=cond)
 
         sum_result = layers.array_read(array=mem_array, i=i)
-        loss = layers.mean(x=sum_result)
+        loss = layers.mean(sum_result)
 
         append_backward(loss)