diff --git a/cmake/inference_lib.cmake b/cmake/inference_lib.cmake
index c6979713231f631f8757e4139d6f685d4554b54e..2fe474fd2eb80cbb0d9ffdcaeed01dbd956d3c58 100644
--- a/cmake/inference_lib.cmake
+++ b/cmake/inference_lib.cmake
@@ -138,25 +138,22 @@ 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}/api.h ${PADDLE_BINARY_DIR}/paddle/fluid/inference/api/libpaddle_inference_api*
+  DSTS ${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/fluid/inference/api/demo_ci/CMakeLists.txt b/paddle/fluid/inference/api/demo_ci/CMakeLists.txt
index 33a0dfaf88b692f5d290920d97a650ad67f3dbca..7f9bb4b33e97b5ea37e9216b00ce0c82ca3ce230 100644
--- a/paddle/fluid/inference/api/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
       ${PADDLE_LIB}/paddle/fluid/inference/libpaddle_fluid.so)
 endif()
 set(EXTERNAL_LIB "-lrt -ldl -lpthread")
diff --git a/paddle/fluid/inference/api/demo_ci/run.sh b/paddle/fluid/inference/api/demo_ci/run.sh
index 87690ac9603e1e54a46a761149798b75ea8a7990..3e829dd726b132844a45427b7b0b39eedf197496 100755
--- a/paddle/fluid/inference/api/demo_ci/run.sh
+++ b/paddle/fluid/inference/api/demo_ci/run.sh
@@ -64,7 +64,7 @@ for WITH_STATIC_LIB in ON OFF; do
     -DWITH_GPU=$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/fluid/inference/api/demo_ci/simple_on_word2vec.cc b/paddle/fluid/inference/api/demo_ci/simple_on_word2vec.cc
index bb4201b687439d21ba3ece6b0ad3a7aa7d579fb4..ecf4a3a7985e30b4c9d376120dc7ce707eaeb49b 100644
--- a/paddle/fluid/inference/api/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/api.h"
 #include "paddle/fluid/platform/enforce.h"
 
 DEFINE_string(dirname, "", "Directory of the inference model.");
diff --git a/paddle/fluid/inference/api/demo_ci/utils.h b/paddle/fluid/inference/api/demo_ci/utils.h
index 3fcf759ad0c193db70e21eae84e4883445e924b7..7298b969ff1ed1898fb119fbd12e36dcb9de4234 100644
--- a/paddle/fluid/inference/api/demo_ci/utils.h
+++ b/paddle/fluid/inference/api/demo_ci/utils.h
@@ -13,10 +13,10 @@
 // limitations under the License.
 
 #pragma once
+#include <algorithm>
 #include <string>
 #include <vector>
-
-#include "contrib/inference/paddle_inference_api.h"
+#include "paddle/fluid/inference/api.h"
 
 namespace paddle {
 namespace demo {
diff --git a/paddle/scripts/paddle_build.sh b/paddle/scripts/paddle_build.sh
index 4bdd86990f82be96085dca640cd52ffb843b2c32..d8e41268ecbc8d94ae63a46951431f692ab1d79a 100755
--- a/paddle/scripts/paddle_build.sh
+++ b/paddle/scripts/paddle_build.sh
@@ -531,7 +531,7 @@ 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
 }