Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into prefetch_on_server

.pre-commit-config.yaml

+repos:
 -   repo: https://github.com/Lucas-C/pre-commit-hooks.git
     sha: v1.0.1
     hooks:
@@ -25,6 +26,14 @@
         entry: bash ./.clang_format.hook -i
         language: system
         files: \.(c|cc|cxx|cpp|cu|h|hpp|hxx|proto)$
+-   repo: local
+    hooks:
+    -   id: cpplint-cpp-source
+        name: cpplint
+        description: Check C++ code style using cpplint.py.
+        entry: bash ./tools/codestyle/cpplint_pre_commit.hook
+        language: system
+        files: \.(c|cc|cxx|cpp|cu|h|hpp|hxx)$
 -   repo: https://github.com/PaddlePaddle/pre-commit-golang
     sha: 8337620115c25ff8333f1b1a493bd031049bd7c0
     hooks:

doc/CMakeLists.txt

+add_custom_target(paddle_apis ALL
+                  DEPENDS paddle_v2_apis paddle_fluid_apis)
+
+add_custom_target(paddle_docs ALL
+                  DEPENDS paddle_v2_docs paddle_v2_docs_cn
+                  paddle_fluid_docs paddle_fluid_docs_cn)
+
 add_subdirectory(v2)
 add_subdirectory(fluid)

doc/fluid/CMakeLists.txt

@@ -27,6 +27,8 @@ sphinx_add_target(paddle_fluid_docs
                   ${CMAKE_CURRENT_SOURCE_DIR}
                   ${SPHINX_HTML_DIR_EN})
 
+add_dependencies(paddle_fluid_docs gen_proto_py)
+
 # configured documentation tools and intermediate build results
 set(BINARY_BUILD_DIR_CN "${CMAKE_CURRENT_BINARY_DIR}/cn/_build")
 
@@ -47,3 +49,7 @@ sphinx_add_target(paddle_fluid_docs_cn
                   ${SPHINX_CACHE_DIR_CN}
                   ${CMAKE_CURRENT_SOURCE_DIR}
                   ${SPHINX_HTML_DIR_CN})
+
+add_dependencies(paddle_fluid_docs_cn gen_proto_py)
+
+add_subdirectory(api)

doc/fluid/api/CMakeLists.txt

+# configured documentation tools and intermediate build results
+set(BINARY_BUILD_DIR_EN "${CMAKE_CURRENT_BINARY_DIR}/en/_build")
+
+# Sphinx cache with pickled ReST documents
+set(SPHINX_CACHE_DIR_EN "${CMAKE_CURRENT_BINARY_DIR}/en/_doctrees")
+
+# HTML output director
+set(SPHINX_HTML_DIR_EN "${CMAKE_CURRENT_BINARY_DIR}/en/html")
+
+configure_file(
+    "${CMAKE_CURRENT_SOURCE_DIR}/../../templates/conf.py.en.in"
+    "${BINARY_BUILD_DIR_EN}/conf.py"
+    @ONLY)
+
+sphinx_add_target(paddle_fluid_apis
+                  html
+                  ${BINARY_BUILD_DIR_EN}
+                  ${SPHINX_CACHE_DIR_EN}
+                  ${CMAKE_CURRENT_SOURCE_DIR}
+                  ${SPHINX_HTML_DIR_EN})
+
+add_dependencies(paddle_fluid_apis  gen_proto_py framework_py_proto copy_paddle_pybind)

doc/v2/CMakeLists.txt

@@ -20,13 +20,15 @@ configure_file(
     "${BINARY_BUILD_DIR_EN}/conf.py"
     @ONLY)
 
-sphinx_add_target(paddle_docs
+sphinx_add_target(paddle_v2_docs
                   html
                   ${BINARY_BUILD_DIR_EN}
                   ${SPHINX_CACHE_DIR_EN}
                   ${CMAKE_CURRENT_SOURCE_DIR}
                   ${SPHINX_HTML_DIR_EN})
 
+add_dependencies(paddle_v2_docs gen_proto_py)
+
 # configured documentation tools and intermediate build results
 set(BINARY_BUILD_DIR_CN "${CMAKE_CURRENT_BINARY_DIR}/cn/_build")
 
@@ -41,11 +43,13 @@ configure_file(
     "${BINARY_BUILD_DIR_CN}/conf.py"
     @ONLY)
 
-sphinx_add_target(paddle_docs_cn
+sphinx_add_target(paddle_v2_docs_cn
                   html
                   ${BINARY_BUILD_DIR_CN}
                   ${SPHINX_CACHE_DIR_CN}
                   ${CMAKE_CURRENT_SOURCE_DIR}
                   ${SPHINX_HTML_DIR_CN})
 
+add_dependencies(paddle_v2_docs_cn gen_proto_py)
+
 add_subdirectory(api)

doc/v2/api/CMakeLists.txt

@@ -12,9 +12,11 @@ configure_file(
     "${BINARY_BUILD_DIR_EN}/conf.py"
     @ONLY)
 
-sphinx_add_target(paddle_api_docs
+sphinx_add_target(paddle_v2_apis
                   html
                   ${BINARY_BUILD_DIR_EN}
                   ${SPHINX_CACHE_DIR_EN}
                   ${CMAKE_CURRENT_SOURCE_DIR}
                   ${SPHINX_HTML_DIR_EN})
+
+add_dependencies(paddle_v2_apis  gen_proto_py framework_py_proto copy_paddle_pybind)

doc/v2/howto/cmd_parameter/index_en.rst

@@ -2,10 +2,25 @@
 
 Set Command-line Parameters
 ===========================
+The implementation of deep learning algorithms has a variety of characteristics, such as running environment, running stage, structure of the model and the traning strategy. PaddlePaddle supports the user to set various command-line parameters flexibly, which helps to achieve control of the model training or prediction process.
+
+In this part, we take several actual scenarios as an example, and the use of some command-line parameters is displayed:
 
 ..  toctree::
   :maxdepth: 1
 
   use_case_en.md
+
+Then, we summarize and classify the use of all command-line parameters:
+
+..  toctree::
+  :maxdepth: 1
+
   arguments_en.md
+
+Finally, the detailed descriptions are given, and we try to explain the propeties and significance of these command-line parameters in detail:
+
+..  toctree::
+  :maxdepth: 1
+
   detail_introduction_en.md

paddle/fluid/operators/detail/grpc_server.cc

@@ -224,13 +224,13 @@ void AsyncGRPCServer::ShutdownQueue() {
   std::unique_lock<std::mutex> lock(cq_mutex_);
   cq_send_->Shutdown();
   cq_get_->Shutdown();
-  is_shut_down_ = true;
 }
 
 // This URL explains why shutdown is complicate:
 void AsyncGRPCServer::ShutDown() {
-  server_->Shutdown();
+  is_shut_down_ = true;
   ShutdownQueue();
+  server_->Shutdown();
 }
 
 void AsyncGRPCServer::TryToRegisterNewSendOne() {
@@ -276,14 +276,14 @@ void AsyncGRPCServer::HandleRequest(::grpc::ServerCompletionQueue* cq,
   bool ok = false;
   while (true) {
     if (!cq->Next(&tag, &ok)) {
-      LOG(INFO) << cq_name << " get CompletionQueue shutdown!";
+      LOG(INFO) << cq_name << " CompletionQueue shutdown!";
       break;
     }
 
     PADDLE_ENFORCE(tag);
     // FIXME(typhoonzero): de-couple the barriers with recv_op
-    if (cq_name == "cq_get") WaitCond(1);
-    if (cq_name == "cq_send") WaitCond(0);
+    if (!is_shut_down_ && cq_name == "cq_get") WaitCond(1);
+    if (!is_shut_down_ && cq_name == "cq_send") WaitCond(0);
 
     RequestBase* base = (RequestBase*)tag;
     // reference:

paddle/fluid/operators/dropout_op.h

@@ -11,9 +11,10 @@ distributed under the License is distributed on an "AS IS" BASIS,
 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. */
-
 #pragma once
+
 #include <random>
+
 #include "paddle/fluid/framework/eigen.h"
 #include "paddle/fluid/framework/op_registry.h"
 

paddle/fluid/operators/dropout_op_test.cc

@@ -13,8 +13,10 @@ See the License for the specific language governing permissions and
 limitations under the License. */
 
 #include <unistd.h>
+
 #include <string>
-#include <thread>
+#include <thread>  // NOLINT
+#include <vector>
 
 #include "gtest/gtest.h"
 #include "paddle/fluid/framework/op_registry.h"
@@ -30,9 +32,9 @@ namespace m = paddle::operators::math;
 
 USE_OP(dropout);
 
-void Compare(f::Scope& scope, p::DeviceContext& ctx) {
+void Compare(f::Scope* scope, const p::DeviceContext& ctx) {
   // init
-  auto var = scope.Var("X");
+  auto var = scope->Var("X");
   auto tensor = var->GetMutable<f::LoDTensor>();
   tensor->Resize({10, 10});
 
@@ -44,12 +46,12 @@ void Compare(f::Scope& scope, p::DeviceContext& ctx) {
   TensorFromVector(init, ctx, tensor);
 
   auto place = ctx.GetPlace();
-  auto out_var = scope.Var("Out");
+  auto out_var = scope->Var("Out");
   auto out_tensor = out_var->GetMutable<f::LoDTensor>();
   out_tensor->Resize({10, 10});
   out_tensor->mutable_data<float>(place);  // allocate
 
-  auto mask_var = scope.Var("Mask");
+  auto mask_var = scope->Var("Mask");
   auto mask_tensor = mask_var->GetMutable<f::LoDTensor>();
   mask_tensor->Resize({10, 10});
   mask_tensor->mutable_data<float>(place);  // allocate
@@ -63,7 +65,7 @@ void Compare(f::Scope& scope, p::DeviceContext& ctx) {
   auto dropout_op = f::OpRegistry::CreateOp(
       "dropout", {{"X", {"X"}}}, {{"Out", {"Out"}}, {"Mask", {"Mask"}}}, attrs);
 
-  dropout_op->Run(scope, place);
+  dropout_op->Run(*scope, place);
 
   std::vector<float> out_vec;
   TensorToVector(*out_tensor, ctx, &out_vec);
@@ -81,6 +83,11 @@ void Compare(f::Scope& scope, p::DeviceContext& ctx) {
   }
 }
 
+// TODO(wyi): Due to
+// https://github.com/PaddlePaddle/Paddle/issues/9507, I temporarily
+// disable this test to remove the prevention of the merge of
+// unrelated PRs.
+/*
 TEST(Dropout, CPUDense) {
   f::Scope scope;
   p::CPUPlace place;
@@ -94,3 +101,4 @@ TEST(Dropout, GPUDense) {
   p::CUDADeviceContext ctx(place);
   Compare(scope, ctx);
 }
+*/

paddle/fluid/operators/listen_and_serv_op.cc

@@ -88,7 +88,6 @@ class ListenAndServOp : public framework::OperatorBase {
 
   void Stop() override {
     rpc_service_->Push(LISTEN_TERMINATE_MESSAGE);
-    rpc_service_->ShutDown();
     server_thread_->join();
   }
 

paddle/fluid/operators/lrn_op.cc

@@ -214,7 +214,10 @@ class LRNOpMaker : public framework::OpProtoAndCheckerMaker {
         "Defaults to \"NHWC\". Specify the data format of the output data, "
         "the input will be transformed automatically. ")
         .SetDefault("AnyLayout");
-    AddAttr<bool>("is_test", "").SetDefault(false);
+    AddAttr<bool>("is_test",
+                  "Turns on memory optimization that optimizes away "
+                  "unnecessary memory allocations. Used by MKLDNN.")
+        .SetDefault(false);
 
     AddComment(R"DOC(
 Local Response Normalization Operator.

paddle/fluid/operators/lrn_op.h

@@ -121,6 +121,10 @@ class LRNGradKernel : public framework::OpKernel<T> {
     T alpha = ctx.Attr<T>("alpha");
     T beta = ctx.Attr<T>("beta");
 
+    PADDLE_ENFORCE(
+        !ctx.Attr<bool>("is_test"),
+        "is_test attribute should be set to False in training phase.");
+
     LRNGradFunctor<DeviceContext, T> f;
     f(ctx, x, out, mid, x_g, out_g, N, C, H, W, n, alpha, beta);
   }

paddle/fluid/operators/send_recv_op_test.cc

@@ -122,7 +122,8 @@ void StartServerNet(bool is_sparse) {
 
   // sub program run in listen_and_serv_op, for simple test we use sum
   f::ProgramDesc program;
-  f::BlockDesc *optimize_block = program.MutableBlock(0);
+  const auto &root_block = program.Block(0);
+  auto *optimize_block = program.AppendBlock(root_block);
   // X for server side tensors, RX for received tensers, must be of same shape.
   AddOp("sum", {{"X", {"x0", "x1"}}}, {{"Out", {"Out"}}}, {}, optimize_block);
 

paddle/scripts/docker/build.sh

@@ -125,9 +125,8 @@ EOF
             -DWITH_AVX=${WITH_AVX:-ON} \
             -DWITH_SWIG_PY=ON \
             -DWITH_STYLE_CHECK=OFF
-        make -j `nproc` gen_proto_py framework_py_proto
-        make -j `nproc` copy_paddle_pybind
-        make -j `nproc` paddle_docs paddle_docs_cn paddle_api_docs
+
+        make -j `nproc` paddle_docs paddle_apis
         popd
     fi
 

paddle/scripts/travis/build_doc.sh

@@ -7,9 +7,8 @@ cd $TRAVIS_BUILD_DIR/build
 
 # Compile Documentation only.
 cmake .. -DCMAKE_BUILD_TYPE=Release -DWITH_GPU=OFF -DWITH_MKL=OFF -DWITH_DOC=ON -DWITH_STYLE_CHECK=OFF
-make -j `nproc` gen_proto_py framework_py_proto
-make -j `nproc` copy_paddle_pybind
-make -j `nproc` paddle_docs paddle_docs_cn paddle_api_docs
+
+make -j `nproc` paddle_docs paddle_apis
 
 # check websites for broken links
 linkchecker doc/v2/en/html/index.html

tools/codestyle/cpplint_pre_commit.hook

+#!/bin/bash
+
+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
+    cpplint $file;
+    TOTAL_ERRORS=$(expr $TOTAL_ERRORS + $?);
+done
+
+exit $TOTAL_ERRORS
+