diff --git a/CMakeLists.txt b/CMakeLists.txt
index 4861e8065a8a70bcc3b859b121cf404461f47ba7..0ab80987b3ad6c4793ceeac1bf3808d2e87fbd5b 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -282,7 +282,3 @@ if(WITH_DOC)
     find_python_module(recommonmark REQUIRED)
     add_subdirectory(doc)
 endif()
-
-if (WITH_CONTRIB)
-    add_subdirectory(paddle/contrib)
-endif()
diff --git a/cmake/inference_lib.cmake b/cmake/inference_lib.cmake
index c6979713231f631f8757e4139d6f685d4554b54e..e2c58cd56055455e7fedc598ca8f56183d4b51dc 100644
--- a/cmake/inference_lib.cmake
+++ b/cmake/inference_lib.cmake
@@ -138,25 +138,24 @@ copy(memory_lib
 
 set(inference_deps paddle_fluid_shared paddle_fluid)
 
-if(WITH_CONTRIB)
-    message(STATUS "installing contrib")
-    set(contrib_dst_dir "${FLUID_INSTALL_DIR}/contrib/inference")
-    if (WITH_ANAKIN AND WITH_GPU)
-        copy(contrib_anakin_inference_lib DEPS paddle_inference_api inference_anakin_api
-            SRCS
-            ${PADDLE_BINARY_DIR}/paddle/contrib/inference/libinference_anakin_api* # compiled anakin api
-            ${PADDLE_BINARY_DIR}/third_party/install/anakin/*.tar.gz # anakin release
-            DSTS ${contrib_dst_dir}/anakin ${contrib_dst_dir}/anakin)
-        list(APPEND inference_deps contrib_anakin_inference_lib)
-   endif()
-
-  copy(contrib_inference_lib DEPS paddle_inference_api paddle_inference_api_shared
-        SRCS ${PADDLE_SOURCE_DIR}/paddle/contrib/inference/paddle_inference_api.h
-        ${PADDLE_BINARY_DIR}/paddle/contrib/inference/libpaddle_inference_api*
-        DSTS ${contrib_dst_dir} ${contrib_dst_dir})
-  list(APPEND inference_deps contrib_inference_lib)
+set(module "inference/api")
+if (WITH_ANAKIN AND WITH_GPU)
+    copy(anakin_inference_lib DEPS paddle_inference_api inference_anakin_api
+        SRCS
+        ${PADDLE_BINARY_DIR}/paddle/fluid/inference/api/libinference_anakin_api* # compiled anakin api
+        ${PADDLE_BINARY_DIR}/third_party/install/anakin/*.tar.gz # anakin release
+        DSTS ${dst_dir}/inference/anakin ${dst_dir}/inference/anakin)
+     list(APPEND inference_deps anakin_inference_lib)
 endif()
 
+copy(inference_api_lib DEPS paddle_inference_api paddle_inference_api_shared
+  SRCS ${src_dir}/${module}/paddle_inference_api.h 
+       ${src_dir}/${module}/demo_ci
+       ${PADDLE_BINARY_DIR}/paddle/fluid/inference/api/libpaddle_inference_api*
+  DSTS ${dst_dir}/inference ${dst_dir}/inference ${dst_dir}/inference
+)
+list(APPEND inference_deps inference_api_lib)
+
 set(module "inference")
 copy(inference_lib DEPS ${inference_deps}
   SRCS ${src_dir}/${module}/*.h ${PADDLE_BINARY_DIR}/paddle/fluid/inference/libpaddle_fluid.*
diff --git a/paddle/contrib/CMakeLists.txt b/paddle/contrib/CMakeLists.txt
deleted file mode 100644
index 4b19256ef4533a09162edf907f6cd51146517e46..0000000000000000000000000000000000000000
--- a/paddle/contrib/CMakeLists.txt
+++ /dev/null
@@ -1,16 +0,0 @@
-# Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved.
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-# http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# 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.
-#
-
-add_subdirectory(inference)
diff --git a/paddle/fluid/inference/CMakeLists.txt b/paddle/fluid/inference/CMakeLists.txt
index 86643b9aa111a9d73c184097e0c43bbec30c6b31..c9eff0fc28c5ff52f902d3d5a0ebb37fa7619e9c 100644
--- a/paddle/fluid/inference/CMakeLists.txt
+++ b/paddle/fluid/inference/CMakeLists.txt
@@ -5,7 +5,7 @@ if (TENSORRT_FOUND)
   add_subdirectory(tensorrt)
 endif()
 
-set(FLUID_CORE_MODULES proto_desc memory lod_tensor executor )
+set(FLUID_CORE_MODULES proto_desc memory lod_tensor executor)
 
 # TODO(panyx0718): Should this be called paddle_fluid_inference_api_internal?
 cc_library(paddle_fluid_api
@@ -38,3 +38,4 @@ if(WITH_TESTING)
   # both tests/book and analysis depends the models that generated by python/paddle/fluid/tests/book
   add_subdirectory(tests/book)
 endif()
+add_subdirectory(api)
diff --git a/paddle/contrib/inference/CMakeLists.txt b/paddle/fluid/inference/api/CMakeLists.txt
similarity index 84%
rename from paddle/contrib/inference/CMakeLists.txt
rename to paddle/fluid/inference/api/CMakeLists.txt
index df470a0752121e44f4305fd3609c5c41985dfaf7..9d63d08dedf6a1bcdacc51bb83d2ed261bca4117 100644
--- a/paddle/contrib/inference/CMakeLists.txt
+++ b/paddle/fluid/inference/api/CMakeLists.txt
@@ -43,21 +43,21 @@ function(inference_api_test TARGET_NAME)
 endfunction(inference_api_test)
 
 cc_library(paddle_inference_api
-    SRCS paddle_inference_api.cc paddle_inference_api_impl.cc
+    SRCS api.cc api_impl.cc
     DEPS ${FLUID_CORE_MODULES} ${GLOB_OP_LIB})
 if(NOT APPLE)
-  set(LINK_FLAGS "-Wl,--retain-symbols-file ${CMAKE_CURRENT_SOURCE_DIR}/paddle_inference_api.sym")
+  set(LINK_FLAGS "-Wl,--retain-symbols-file ${CMAKE_CURRENT_SOURCE_DIR}/api.sym")
   set_target_properties(paddle_inference_api PROPERTIES LINK_FLAGS "${LINK_FLAGS}")
 endif()
 
 # Here the shared library doesn't depend on other fluid libraries, or double free will occur.
 cc_library(paddle_inference_api_shared SHARED
-    SRCS paddle_inference_api.cc paddle_inference_api_impl.cc)
+    SRCS api.cc api_impl.cc)
 add_dependencies(paddle_inference_api_shared ${FLUID_CORE_MODULES} ${GLOB_OP_LIB})
 set_target_properties(paddle_inference_api_shared PROPERTIES OUTPUT_NAME paddle_inference_api)
 
 if(NOT APPLE)
-  set(LINK_FLAGS "-Wl,--version-script ${CMAKE_CURRENT_SOURCE_DIR}/paddle_inference_api.map")
+  set(LINK_FLAGS "-Wl,--version-script ${CMAKE_CURRENT_SOURCE_DIR}/api.map")
   set_target_properties(paddle_inference_api_shared PROPERTIES LINK_FLAGS "${LINK_FLAGS}")
   FILE(WRITE ${CMAKE_CURRENT_BINARY_DIR}/check_symbol.cmake
     "execute_process(COMMAND bash -c \"${CMAKE_CURRENT_SOURCE_DIR}/check_symbol.sh"
@@ -73,32 +73,32 @@ if(NOT APPLE)
 endif()
 
 cc_test(test_paddle_inference_api
-        SRCS test_paddle_inference_api.cc
+        SRCS test_api.cc
         DEPS paddle_inference_api)
 
-inference_api_test(test_paddle_inference_api_impl
+inference_api_test(test_api_impl
                     ARGS test_word2vec test_image_classification)
 
 if(WITH_GPU AND TENSORRT_FOUND)
 cc_library(paddle_inference_tensorrt_subgraph_engine
-        SRCS paddle_inference_api_tensorrt_subgraph_engine.cc
-        DEPS paddle_inference_api analysis tensorrt_engine paddle_inference_api paddle_fluid_api)
+        SRCS api_tensorrt_subgraph_engine.cc
+        DEPS paddle_inference_api analysis tensorrt_engine paddle_fluid_api)
 
-inference_api_test(test_paddle_inference_api_tensorrt_subgraph_engine ARGS test_word2vec)
+inference_api_test(test_api_tensorrt_subgraph_engine ARGS test_word2vec)
 endif()
 
 if (WITH_ANAKIN) # only needed in CI
     # Due to Anakin do not have official library releases and the versions of protobuf and cuda do not match Paddle's,
     # so anakin library will not be merged to our official inference library. To use anakin prediction API, one need to
     # compile the libinference_anakin_api.a and compile with anakin.so.
-    nv_library(inference_anakin_api SRCS paddle_inference_api.cc paddle_inference_api_anakin_engine.cc)
-    nv_library(inference_anakin_api_shared SHARED SRCS paddle_inference_api.cc paddle_inference_api_anakin_engine.cc)
+    nv_library(inference_anakin_api SRCS api.cc api_anakin_engine.cc)
+    nv_library(inference_anakin_api_shared SHARED SRCS api.cc api_anakin_engine.cc)
     target_compile_options(inference_anakin_api BEFORE PUBLIC ${ANAKIN_COMPILE_EXTRA_FLAGS})
     target_compile_options(inference_anakin_api_shared BEFORE PUBLIC ${ANAKIN_COMPILE_EXTRA_FLAGS})
     target_link_libraries(inference_anakin_api anakin anakin_saber_common)
     target_link_libraries(inference_anakin_api_shared anakin anakin_saber_common)
     if (WITH_TESTING)
-        cc_test(inference_anakin_test SRCS paddle_inference_api_anakin_engine_tester.cc
+        cc_test(inference_anakin_test SRCS api_anakin_engine_tester.cc
                                   ARGS --model=${ANAKIN_INSTALL_DIR}/mobilenet_v2.anakin.bin
                                   DEPS inference_anakin_api)
         target_compile_options(inference_anakin_test BEFORE PUBLIC ${ANAKIN_COMPILE_EXTRA_FLAGS})
diff --git a/paddle/contrib/inference/README.md b/paddle/fluid/inference/api/README.md
similarity index 100%
rename from paddle/contrib/inference/README.md
rename to paddle/fluid/inference/api/README.md
diff --git a/paddle/contrib/inference/paddle_inference_api.cc b/paddle/fluid/inference/api/api.cc
similarity index 96%
rename from paddle/contrib/inference/paddle_inference_api.cc
rename to paddle/fluid/inference/api/api.cc
index 4fe198ad7d4a752882965e9e7fc460741de53d22..e74f23ff969f5a8f58a71da337c16dcbc14f10c0 100644
--- a/paddle/contrib/inference/paddle_inference_api.cc
+++ b/paddle/fluid/inference/api/api.cc
@@ -12,7 +12,7 @@ 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/contrib/inference/paddle_inference_api.h"
+#include "paddle/fluid/inference/api/paddle_inference_api.h"
 
 namespace paddle {
 
diff --git a/paddle/contrib/inference/paddle_inference_api.map b/paddle/fluid/inference/api/api.map
similarity index 100%
rename from paddle/contrib/inference/paddle_inference_api.map
rename to paddle/fluid/inference/api/api.map
diff --git a/paddle/contrib/inference/paddle_inference_api.sym b/paddle/fluid/inference/api/api.sym
similarity index 100%
rename from paddle/contrib/inference/paddle_inference_api.sym
rename to paddle/fluid/inference/api/api.sym
diff --git a/paddle/contrib/inference/paddle_inference_api_anakin_engine.cc b/paddle/fluid/inference/api/api_anakin_engine.cc
similarity index 90%
rename from paddle/contrib/inference/paddle_inference_api_anakin_engine.cc
rename to paddle/fluid/inference/api/api_anakin_engine.cc
index daf0b56825498d48ca48d1d5b0ea6a3d7314bfc4..f6f3cb335897b02905e24c229b92f3940a37dbf8 100644
--- a/paddle/contrib/inference/paddle_inference_api_anakin_engine.cc
+++ b/paddle/fluid/inference/api/api_anakin_engine.cc
@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/contrib/inference/paddle_inference_api_anakin_engine.h"
+#include "paddle/fluid/inference/api/api_anakin_engine.h"
 #include <cuda.h>
+#include <vector>
 
 namespace paddle {
 
@@ -47,8 +48,7 @@ bool PaddleInferenceAnakinPredictor::Run(
     }
     auto d_tensor_in_p = executor_.get_in(input.name);
     float *d_data_p = d_tensor_in_p->mutable_data();
-    if (cudaMemcpy(d_data_p,
-                   static_cast<float *>(input.data.data()),
+    if (cudaMemcpy(d_data_p, static_cast<float *>(input.data.data()),
                    d_tensor_in_p->valid_size() * sizeof(float),
                    cudaMemcpyHostToDevice) != 0) {
       LOG(ERROR) << "copy data from CPU to GPU error";
@@ -70,8 +70,7 @@ bool PaddleInferenceAnakinPredictor::Run(
       output.data.Resize(tensor->valid_size() * sizeof(float));
     }
     // Copy data from GPU -> CPU
-    if (cudaMemcpy(output.data.data(),
-                   tensor->mutable_data(),
+    if (cudaMemcpy(output.data.data(), tensor->mutable_data(),
                    tensor->valid_size() * sizeof(float),
                    cudaMemcpyDeviceToHost) != 0) {
       LOG(ERROR) << "copy data from GPU to CPU error";
@@ -106,13 +105,12 @@ std::unique_ptr<PaddlePredictor> PaddleInferenceAnakinPredictor::Clone() {
 
 // A factory to help create difference predictor.
 template <>
-std::unique_ptr<PaddlePredictor>
-CreatePaddlePredictor<AnakinConfig, PaddleEngineKind::kAnakin>(
-    const AnakinConfig &config) {
+std::unique_ptr<PaddlePredictor> CreatePaddlePredictor<
+    AnakinConfig, PaddleEngineKind::kAnakin>(const AnakinConfig &config) {
   VLOG(3) << "Anakin Predictor create.";
   std::unique_ptr<PaddlePredictor> x(
       new PaddleInferenceAnakinPredictor(config));
   return x;
-};
+}
 
 }  // namespace paddle
diff --git a/paddle/contrib/inference/paddle_inference_api_anakin_engine.h b/paddle/fluid/inference/api/api_anakin_engine.h
similarity index 88%
rename from paddle/contrib/inference/paddle_inference_api_anakin_engine.h
rename to paddle/fluid/inference/api/api_anakin_engine.h
index 212ba41cdf8ff2feccb6b6498f9679d76a2efe7c..85ca83cd00756cca04d7b92437e9955d8ab297e7 100644
--- a/paddle/contrib/inference/paddle_inference_api_anakin_engine.h
+++ b/paddle/fluid/inference/api/api_anakin_engine.h
@@ -19,7 +19,8 @@ limitations under the License. */
 
 #pragma once
 
-#include "paddle/contrib/inference/paddle_inference_api.h"
+#include <vector>
+#include "paddle/fluid/inference/api/paddle_inference_api.h"
 
 // from anakin
 #include "framework/core/net/net.h"
@@ -31,7 +32,7 @@ class PaddleInferenceAnakinPredictor : public PaddlePredictor {
  public:
   PaddleInferenceAnakinPredictor() {}
 
-  PaddleInferenceAnakinPredictor(const AnakinConfig& config);
+  explicit PaddleInferenceAnakinPredictor(const AnakinConfig& config);
 
   // NOTE Unlike the native engine, the buffers of anakin engine's output_data
   // should be allocated first.
@@ -48,8 +49,7 @@ class PaddleInferenceAnakinPredictor : public PaddlePredictor {
  private:
   bool Init(const AnakinConfig& config);
 
-  anakin::graph::Graph<anakin::NV,
-                       anakin::saber::AK_FLOAT,
+  anakin::graph::Graph<anakin::NV, anakin::saber::AK_FLOAT,
                        anakin::Precision::FP32>
       graph_;
   anakin::Net<anakin::NV, anakin::saber::AK_FLOAT, anakin::Precision::FP32>
diff --git a/paddle/contrib/inference/paddle_inference_api_anakin_engine_tester.cc b/paddle/fluid/inference/api/api_anakin_engine_tester.cc
similarity index 97%
rename from paddle/contrib/inference/paddle_inference_api_anakin_engine_tester.cc
rename to paddle/fluid/inference/api/api_anakin_engine_tester.cc
index f92e9d4190412f5847e353ef1dc0324cad668c9a..d6d631bfbad4278fe99e4553a410a9d9162dcc7b 100644
--- a/paddle/contrib/inference/paddle_inference_api_anakin_engine_tester.cc
+++ b/paddle/fluid/inference/api/api_anakin_engine_tester.cc
@@ -16,7 +16,7 @@ limitations under the License. */
 #include <glog/logging.h>
 #include <gtest/gtest.h>
 
-#include "paddle/contrib/inference/paddle_inference_api.h"
+#include "paddle/fluid/inference/api/paddle_inference_api.h"
 
 DEFINE_string(model, "", "Directory of the inference model.");
 
diff --git a/paddle/contrib/inference/paddle_inference_api_impl.cc b/paddle/fluid/inference/api/api_impl.cc
similarity index 92%
rename from paddle/contrib/inference/paddle_inference_api_impl.cc
rename to paddle/fluid/inference/api/api_impl.cc
index b1e5b875981e0142f6970cf6864b7b598743654b..9d9e126e134deafc18f10f7299e81d92702e9ca0 100644
--- a/paddle/contrib/inference/paddle_inference_api_impl.cc
+++ b/paddle/fluid/inference/api/api_impl.cc
@@ -21,7 +21,7 @@ limitations under the License. */
 #include <utility>
 #include <vector>
 
-#include "paddle/contrib/inference/paddle_inference_api_impl.h"
+#include "paddle/fluid/inference/api/api_impl.h"
 
 namespace paddle {
 namespace {
@@ -77,8 +77,8 @@ bool NativePaddlePredictor::Init(
   if (!config_.model_dir.empty()) {
     // Parameters are saved in separate files sited in
     // the specified `dirname`.
-    inference_program_ = paddle::inference::Load(
-        executor_.get(), scope_.get(), config_.model_dir);
+    inference_program_ = paddle::inference::Load(executor_.get(), scope_.get(),
+                                                 config_.model_dir);
   } else if (!config_.prog_file.empty() && !config_.param_file.empty()) {
     // All parameters are saved in a single file.
     // The file names should be consistent with that used
@@ -91,8 +91,8 @@ bool NativePaddlePredictor::Init(
   }
 
   ctx_ = executor_->Prepare(*inference_program_, 0);
-  executor_->CreateVariables(
-      *inference_program_, sub_scope_ ? sub_scope_ : scope_.get(), 0);
+  executor_->CreateVariables(*inference_program_,
+                             sub_scope_ ? sub_scope_ : scope_.get(), 0);
 
   // Get the feed_target_names and fetch_target_names
   feed_target_names_ = inference_program_->GetFeedTargetNames();
@@ -105,7 +105,7 @@ NativePaddlePredictor::~NativePaddlePredictor() {
     PADDLE_ENFORCE_NOT_NULL(scope_, "Should have parent scope!");
     scope_->DeleteScope(sub_scope_);
   }
-};
+}
 
 bool NativePaddlePredictor::Run(const std::vector<PaddleTensor> &inputs,
                                 std::vector<PaddleTensor> *output_data) {
@@ -134,10 +134,8 @@ bool NativePaddlePredictor::Run(const std::vector<PaddleTensor> &inputs,
   // if share variables, we need not create variables
   VLOG(4) << "Run prepared context";
   executor_->RunPreparedContext(
-      ctx_.get(),
-      sub_scope_ != nullptr ? sub_scope_ : scope_.get(),
-      &feed_targets,
-      &fetch_targets,
+      ctx_.get(), sub_scope_ != nullptr ? sub_scope_ : scope_.get(),
+      &feed_targets, &fetch_targets,
       false /* don't create variable eatch time */);
   VLOG(4) << "Finish prepared context";
   if (!GetFetch(fetchs, output_data)) {
@@ -181,8 +179,7 @@ bool NativePaddlePredictor::SetFeed(const std::vector<PaddleTensor> &inputs,
     }
 
     // TODO(panyx0718): Init LoDTensor from existing memcpy to save a copy.
-    std::memcpy(static_cast<void *>(input_ptr),
-                inputs[i].data.data(),
+    std::memcpy(static_cast<void *>(input_ptr), inputs[i].data.data(),
                 inputs[i].data.length());
     feeds->push_back(input);
   }
@@ -232,8 +229,7 @@ bool NativePaddlePredictor::GetFetch(
         size_t start = lod[0][j - 1] * common_dim;
         size_t end = lod[0][j] * common_dim;
         if (end > start) {
-          std::copy(output_ptr + start,
-                    output_ptr + end,
+          std::copy(output_ptr + start, output_ptr + end,
                     data.begin() + (j - 1) * max_dim * common_dim);
         }
       }
@@ -257,15 +253,13 @@ bool NativePaddlePredictor::GetFetch(
 }
 
 template <>
-std::unique_ptr<PaddlePredictor>
-CreatePaddlePredictor<NativeConfig, PaddleEngineKind::kNative>(
-    const NativeConfig &config) {
+std::unique_ptr<PaddlePredictor> CreatePaddlePredictor<
+    NativeConfig, PaddleEngineKind::kNative>(const NativeConfig &config) {
   VLOG(3) << "create NativePaddlePredictor";
   if (config.use_gpu) {
     // 1. GPU memeroy
     PADDLE_ENFORCE_GT(
-        config.fraction_of_gpu_memory,
-        0.f,
+        config.fraction_of_gpu_memory, 0.f,
         "fraction_of_gpu_memory in the config should be set to range (0., 1.]");
     PADDLE_ENFORCE_GE(config.device, 0, "Invalid device id %d", config.device);
     std::vector<std::string> flags;
diff --git a/paddle/contrib/inference/paddle_inference_api_impl.h b/paddle/fluid/inference/api/api_impl.h
similarity index 97%
rename from paddle/contrib/inference/paddle_inference_api_impl.h
rename to paddle/fluid/inference/api/api_impl.h
index f9ec6f55449fc46b4a44b9563980cb5f8e80a951..92e693578ab657004f3c40c09b979897afea1e1f 100644
--- a/paddle/contrib/inference/paddle_inference_api_impl.h
+++ b/paddle/fluid/inference/api/api_impl.h
@@ -19,7 +19,7 @@
 #include <string>
 #include <vector>
 
-#include "paddle/contrib/inference/paddle_inference_api.h"
+#include "paddle/fluid/inference/api/paddle_inference_api.h"
 
 #include "paddle/fluid/framework/ddim.h"
 #include "paddle/fluid/framework/lod_tensor.h"
diff --git a/paddle/contrib/inference/paddle_inference_api_tensorrt_subgraph_engine.cc b/paddle/fluid/inference/api/api_tensorrt_subgraph_engine.cc
similarity index 93%
rename from paddle/contrib/inference/paddle_inference_api_tensorrt_subgraph_engine.cc
rename to paddle/fluid/inference/api/api_tensorrt_subgraph_engine.cc
index a11396cee91a758e86af2efd9e58b9da68442590..0cdc88fa1eaf3935ce0da143e1e91eb84cd70dcf 100644
--- a/paddle/contrib/inference/paddle_inference_api_tensorrt_subgraph_engine.cc
+++ b/paddle/fluid/inference/api/api_tensorrt_subgraph_engine.cc
@@ -12,9 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/contrib/inference/paddle_inference_api.h"
-#include "paddle/contrib/inference/paddle_inference_api_impl.h"
 #include "paddle/fluid/inference/analysis/analyzer.h"
+#include "paddle/fluid/inference/api/api_impl.h"
+#include "paddle/fluid/inference/api/paddle_inference_api.h"
 #include "paddle/fluid/inference/utils/singleton.h"
 
 namespace paddle {
@@ -77,8 +77,8 @@ class TensorRTSubgraphPredictor : public NativePaddlePredictor {
     ctx_ = executor_->Prepare(*inference_program_, 0);
 
     VLOG(5) << "to create variables";
-    executor_->CreateVariables(
-        *inference_program_, sub_scope_ ? sub_scope_ : scope_.get(), 0);
+    executor_->CreateVariables(*inference_program_,
+                               sub_scope_ ? sub_scope_ : scope_.get(), 0);
 
     // Get the feed_target_names and fetch_target_names
     feed_target_names_ = inference_program_->GetFeedTargetNames();
@@ -98,8 +98,7 @@ CreatePaddlePredictor<TensorRTConfig, PaddleEngineKind::kAutoMixedTensorRT>(
   if (config.use_gpu) {
     // 1. GPU memeroy
     PADDLE_ENFORCE_GT(
-        config.fraction_of_gpu_memory,
-        0.f,
+        config.fraction_of_gpu_memory, 0.f,
         "fraction_of_gpu_memory in the config should be set to range (0., 1.]");
     PADDLE_ENFORCE_GE(config.device, 0, "Invalid device id %d", config.device);
     std::vector<std::string> flags;
diff --git a/paddle/contrib/inference/check_symbol.sh b/paddle/fluid/inference/api/check_symbol.sh
similarity index 100%
rename from paddle/contrib/inference/check_symbol.sh
rename to paddle/fluid/inference/api/check_symbol.sh
diff --git a/paddle/contrib/inference/demo_ci/.gitignore b/paddle/fluid/inference/api/demo_ci/.gitignore
similarity index 100%
rename from paddle/contrib/inference/demo_ci/.gitignore
rename to paddle/fluid/inference/api/demo_ci/.gitignore
diff --git a/paddle/contrib/inference/demo_ci/CMakeLists.txt b/paddle/fluid/inference/api/demo_ci/CMakeLists.txt
similarity index 95%
rename from paddle/contrib/inference/demo_ci/CMakeLists.txt
rename to paddle/fluid/inference/api/demo_ci/CMakeLists.txt
index 33a0dfaf88b692f5d290920d97a650ad67f3dbca..7f9bb4b33e97b5ea37e9216b00ce0c82ca3ce230 100644
--- a/paddle/contrib/inference/demo_ci/CMakeLists.txt
+++ b/paddle/fluid/inference/api/demo_ci/CMakeLists.txt
@@ -55,11 +55,11 @@ endif()
 # Note: libpaddle_inference_api.so/a must put before libpaddle_fluid.so/a
 if(WITH_STATIC_LIB)
   set(DEPS
-      ${PADDLE_LIB}/contrib/inference/libpaddle_inference_api.a
+      ${PADDLE_LIB}/paddle/fluid/inference/libpaddle_inference_api.a
       ${PADDLE_LIB}/paddle/fluid/inference/libpaddle_fluid.a)
 else()
   set(DEPS
-      ${PADDLE_LIB}/contrib/inference/libpaddle_inference_api.so
+      ${PADDLE_LIB}/paddle/fluid/inference/libpaddle_inference_api.so
       ${PADDLE_LIB}/paddle/fluid/inference/libpaddle_fluid.so)
 endif()
 set(EXTERNAL_LIB "-lrt -ldl -lpthread")
diff --git a/paddle/contrib/inference/demo_ci/README.md b/paddle/fluid/inference/api/demo_ci/README.md
similarity index 100%
rename from paddle/contrib/inference/demo_ci/README.md
rename to paddle/fluid/inference/api/demo_ci/README.md
diff --git a/paddle/contrib/inference/demo_ci/run.sh b/paddle/fluid/inference/api/demo_ci/run.sh
similarity index 98%
rename from paddle/contrib/inference/demo_ci/run.sh
rename to paddle/fluid/inference/api/demo_ci/run.sh
index 87690ac9603e1e54a46a761149798b75ea8a7990..3e829dd726b132844a45427b7b0b39eedf197496 100755
--- a/paddle/contrib/inference/demo_ci/run.sh
+++ b/paddle/fluid/inference/api/demo_ci/run.sh
@@ -64,7 +64,7 @@ for WITH_STATIC_LIB in ON OFF; do
     -DWITH_GPU=$TEST_GPU_CPU \
     -DWITH_STATIC_LIB=$WITH_STATIC_LIB
   make -j
-  for use_gpu in false; do
+  for use_gpu in $use_gpu_list; do
     for vis_demo_name in $vis_demo_list; do 
       ./vis_demo \
         --modeldir=../data/$vis_demo_name/model \
diff --git a/paddle/contrib/inference/demo_ci/simple_on_word2vec.cc b/paddle/fluid/inference/api/demo_ci/simple_on_word2vec.cc
similarity index 93%
rename from paddle/contrib/inference/demo_ci/simple_on_word2vec.cc
rename to paddle/fluid/inference/api/demo_ci/simple_on_word2vec.cc
index 9713837f86d40383da946af1681e1945c84336b0..5f96fecf93f7a6c42bc6b9fe4e0d985c626388d7 100644
--- a/paddle/contrib/inference/demo_ci/simple_on_word2vec.cc
+++ b/paddle/fluid/inference/api/demo_ci/simple_on_word2vec.cc
@@ -19,8 +19,8 @@ limitations under the License. */
 #include <gflags/gflags.h>
 #include <glog/logging.h>
 #include <memory>
-#include <thread>
-#include "contrib/inference/paddle_inference_api.h"
+#include <thread>  //NOLINT
+#include "paddle/fluid/inference/paddle_inference_api.h"
 #include "paddle/fluid/platform/enforce.h"
 
 DEFINE_string(dirname, "", "Directory of the inference model.");
@@ -63,8 +63,8 @@ void Main(bool use_gpu) {
     PADDLE_ENFORCE(outputs.size(), 1UL);
     // Check the output buffer size and result of each tid.
     PADDLE_ENFORCE(outputs.front().data.length(), 33168UL);
-    float result[5] = {
-        0.00129761, 0.00151112, 0.000423564, 0.00108815, 0.000932706};
+    float result[5] = {0.00129761, 0.00151112, 0.000423564, 0.00108815,
+                       0.000932706};
     const size_t num_elements = outputs.front().data.length() / sizeof(float);
     // The outputs' buffers are in CPU memory.
     for (size_t i = 0; i < std::min(5UL, num_elements); i++) {
@@ -107,8 +107,8 @@ void MainThreads(int num_threads, bool use_gpu) {
         PADDLE_ENFORCE(outputs.size(), 1UL);
         // Check the output buffer size and result of each tid.
         PADDLE_ENFORCE(outputs.front().data.length(), 33168UL);
-        float result[5] = {
-            0.00129761, 0.00151112, 0.000423564, 0.00108815, 0.000932706};
+        float result[5] = {0.00129761, 0.00151112, 0.000423564, 0.00108815,
+                           0.000932706};
         const size_t num_elements =
             outputs.front().data.length() / sizeof(float);
         // The outputs' buffers are in CPU memory.
diff --git a/paddle/contrib/inference/demo_ci/utils.h b/paddle/fluid/inference/api/demo_ci/utils.h
similarity index 93%
rename from paddle/contrib/inference/demo_ci/utils.h
rename to paddle/fluid/inference/api/demo_ci/utils.h
index 017b39edaf1dd8fb9d3b66f278abcf9ba7991ba9..cb8990671162dff47228736e69617229528cc093 100644
--- a/paddle/contrib/inference/demo_ci/utils.h
+++ b/paddle/fluid/inference/api/demo_ci/utils.h
@@ -13,16 +13,15 @@
 // limitations under the License.
 
 #pragma once
+#include <algorithm>
 #include <string>
 #include <vector>
-
-#include "contrib/inference/paddle_inference_api.h"
+#include "paddle/fluid/inference/paddle_inference_api.h"
 
 namespace paddle {
 namespace demo {
 
-static void split(const std::string& str,
-                  char sep,
+static void split(const std::string& str, char sep,
                   std::vector<std::string>* pieces) {
   pieces->clear();
   if (str.empty()) {
diff --git a/paddle/contrib/inference/demo_ci/vis_demo.cc b/paddle/fluid/inference/api/demo_ci/vis_demo.cc
similarity index 99%
rename from paddle/contrib/inference/demo_ci/vis_demo.cc
rename to paddle/fluid/inference/api/demo_ci/vis_demo.cc
index 1c570a9f658be5641ac2b3a288376bdf19b293a1..0a2a2b713ab21a3124d8a85ba469f64278623ec4 100644
--- a/paddle/contrib/inference/demo_ci/vis_demo.cc
+++ b/paddle/fluid/inference/api/demo_ci/vis_demo.cc
@@ -29,8 +29,7 @@ DECLARE_double(fraction_of_gpu_memory_to_use);
 DEFINE_string(modeldir, "", "Directory of the inference model.");
 DEFINE_string(refer, "", "path to reference result for comparison.");
 DEFINE_string(
-    data,
-    "",
+    data, "",
     "path of data; each line is a record, format is "
     "'<space splitted floats as data>\t<space splitted ints as shape'");
 DEFINE_bool(use_gpu, false, "Whether use gpu.");
diff --git a/paddle/contrib/inference/high_level_api.md b/paddle/fluid/inference/api/high_level_api.md
similarity index 100%
rename from paddle/contrib/inference/high_level_api.md
rename to paddle/fluid/inference/api/high_level_api.md
diff --git a/paddle/contrib/inference/high_level_api_cn.md b/paddle/fluid/inference/api/high_level_api_cn.md
similarity index 100%
rename from paddle/contrib/inference/high_level_api_cn.md
rename to paddle/fluid/inference/api/high_level_api_cn.md
diff --git a/paddle/contrib/inference/paddle_inference_api.h b/paddle/fluid/inference/api/paddle_inference_api.h
similarity index 100%
rename from paddle/contrib/inference/paddle_inference_api.h
rename to paddle/fluid/inference/api/paddle_inference_api.h
diff --git a/paddle/contrib/inference/test_paddle_inference_api.cc b/paddle/fluid/inference/api/test_api.cc
similarity index 96%
rename from paddle/contrib/inference/test_paddle_inference_api.cc
rename to paddle/fluid/inference/api/test_api.cc
index bc7faab6e208a66d7a56e41a56bd743c7644eea2..ac8a21a22be6f27311b8ae2507d04d9d1b510e76 100644
--- a/paddle/contrib/inference/test_paddle_inference_api.cc
+++ b/paddle/fluid/inference/api/test_api.cc
@@ -12,10 +12,9 @@ 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/contrib/inference/paddle_inference_api.h"
-
 #include <glog/logging.h>
 #include <gtest/gtest.h>
+#include "paddle/fluid/inference/api/paddle_inference_api.h"
 
 namespace paddle {
 
diff --git a/paddle/contrib/inference/test_paddle_inference_api_impl.cc b/paddle/fluid/inference/api/test_api_impl.cc
similarity index 98%
rename from paddle/contrib/inference/test_paddle_inference_api_impl.cc
rename to paddle/fluid/inference/api/test_api_impl.cc
index c3649dcb96c77f449d876bef34c4aea7afb31daa..fc1364b80ac1ee2d304eb2fe429eae5f56967516 100644
--- a/paddle/contrib/inference/test_paddle_inference_api_impl.cc
+++ b/paddle/fluid/inference/api/test_api_impl.cc
@@ -15,10 +15,10 @@ limitations under the License. */
 #include <glog/logging.h>
 #include <gtest/gtest.h>
 
-#include <thread>
+#include <thread>  // NOLINT
 
 #include "gflags/gflags.h"
-#include "paddle/contrib/inference/paddle_inference_api_impl.h"
+#include "paddle/fluid/inference/api/api_impl.h"
 #include "paddle/fluid/inference/tests/test_helper.h"
 
 DEFINE_string(dirname, "", "Directory of the inference model.");
@@ -121,8 +121,8 @@ void MainImageClassification(bool use_gpu) {
   // which should be in the range [0.0, 1.0].
   feed_target_shapes[0][0] = batch_size;
   framework::DDim input_dims = framework::make_ddim(feed_target_shapes[0]);
-  SetupTensor<float>(
-      &input, input_dims, static_cast<float>(0), static_cast<float>(1));
+  SetupTensor<float>(&input, input_dims, static_cast<float>(0),
+                     static_cast<float>(1));
   std::vector<framework::LoDTensor*> cpu_feeds;
   cpu_feeds.push_back(&input);
 
diff --git a/paddle/contrib/inference/test_paddle_inference_api_tensorrt_subgraph_engine.cc b/paddle/fluid/inference/api/test_api_tensorrt_subgraph_engine.cc
similarity index 96%
rename from paddle/contrib/inference/test_paddle_inference_api_tensorrt_subgraph_engine.cc
rename to paddle/fluid/inference/api/test_api_tensorrt_subgraph_engine.cc
index b100630dbe412ca811f1a8f2b8191356f5ebec2f..585f6d29376c3341c21ff76361d5335512c1b1b6 100644
--- a/paddle/contrib/inference/test_paddle_inference_api_tensorrt_subgraph_engine.cc
+++ b/paddle/fluid/inference/api/test_api_tensorrt_subgraph_engine.cc
@@ -15,7 +15,7 @@
 #include <gflags/gflags.h>
 #include <glog/logging.h>
 #include <gtest/gtest.h>
-#include "paddle/contrib/inference/paddle_inference_api.h"
+#include "paddle/fluid/inference/api/paddle_inference_api.h"
 
 namespace paddle {
 
@@ -61,4 +61,4 @@ void Main(bool use_gpu) {
 
 TEST(paddle_inference_api_tensorrt_subgraph_engine, main) { Main(true); }
 
-}  // namespace paddle
\ No newline at end of file
+}  // namespace paddle
diff --git a/paddle/scripts/paddle_build.sh b/paddle/scripts/paddle_build.sh
index de915d4e56671dca8a61a2083e0840358f1b66ea..5f2c366efd9b043b551a88a42b39f703bd052b6b 100755
--- a/paddle/scripts/paddle_build.sh
+++ b/paddle/scripts/paddle_build.sh
@@ -545,7 +545,7 @@ function test_fluid_inference_lib() {
     Testing fluid inference library ...
     ========================================
 EOF
-        cd ${PADDLE_ROOT}/paddle/contrib/inference/demo_ci
+        cd ${PADDLE_ROOT}/paddle/fluid/inference/api/demo_ci
         ./run.sh ${PADDLE_ROOT} ${WITH_MKL:-ON} ${WITH_GPU:-OFF}
       fi
 }