diff --git a/README.md b/README.md
index 23138a067edfd6dcd4c13d9fe90397af074e6517..a95cbf30b6f877860ef5e46c0f7007a6bc5f0d0a 100644
--- a/README.md
+++ b/README.md
@@ -5,16 +5,16 @@
 # VisualDL (Visualize the Deep Learning)
 
 ## Introduction
-VisualDL is a deep learning visualization tool that can help design deep learning jobs. 
-It includes features such as scalar, parameter distribution, model structure and image visualization. 
-Currently it is being developed at a high pace. 
+VisualDL is a deep learning visualization tool that can help design deep learning jobs.
+It includes features such as scalar, parameter distribution, model structure and image visualization.
+Currently it is being developed at a high pace.
 New features will be continuously added.
 
-At present, most DNN frameworks use Python as their primary language. VisualDL supports Python by nature. 
+At present, most DNN frameworks use Python as their primary language. VisualDL supports Python by nature.
 Users can get plentiful visualization results by simply add a few lines of Python code into their model before training.
-  
-  
-Besides Python SDK, VisualDL was writen in C++ on the low level. It also provides C++ SDK that 
+
+
+Besides Python SDK, VisualDL was writen in C++ on the low level. It also provides C++ SDK that
 can be integrated into other platforms.  
 
 
@@ -27,8 +27,8 @@ VisualDL now provides 4 components:
 - histogram
 
 ### Graph
-Graph is compatible with ONNX(Open Neural Network Exchange)[https://github.com/onnx/onnx], 
-Cooperated with Python SDK, VisualDL can be compatible with most major DNN frameworks, including 
+Graph is compatible with ONNX(Open Neural Network Exchange)[https://github.com/onnx/onnx],
+Cooperated with Python SDK, VisualDL can be compatible with most major DNN frameworks, including
 PaddlePaddle, PyTorch and MXNet.
 
 <p align="center">
@@ -75,7 +75,7 @@ logger = LogWriter(dir, sync_cycle=10)
 with logger.mode("train"):
     # create a scalar component called 'scalars/scalar0'
     scalar0 = logger.scalar("scalars/scalar0")
-    
+
 
 # add some records during DL model running, lets start from another block.
 with logger.mode("train"):
@@ -98,12 +98,12 @@ namepsace cp = visualdl::components;
 int main() {
   const std::string dir = "./tmp";
   vs::LogWriter logger(dir, 10);
-  
+
   logger.SetMode("train");
   auto tablet = logger.NewTablet("scalars/scalar0");
-  
+
   cp::Scalar<float> scalar0(tablet);
-  
+
   for (int step = 0; step < 1000; step++) {
     float v = (float)std::rand() / RAND_MAX;
     scalar0.AddRecord(step, v);
@@ -131,7 +131,7 @@ Board also supports the parameters below for remote access:
 ### How to install
 ```
 python setup.py bdist_wheel
-pip install --upgrade dist/visualdl-0.0.1-py2-none-any.whl
+pip install --upgrade dist/visualdl-*.whl
 ```
 
 ### Run a demo from scratch
@@ -144,6 +144,6 @@ that will start a server locally.
 
 ### Contribute
 
-VisualDL is initially created by [PaddlePaddle](http://www.paddlepaddle.org/) and 
-[ECharts](http://echarts.baidu.com/). 
+VisualDL is initially created by [PaddlePaddle](http://www.paddlepaddle.org/) and
+[ECharts](http://echarts.baidu.com/).
 We welcome everyone to use, comment and contribute to Visual DL :)
diff --git a/build.sh b/build.sh
index 426c5aff590c9cb09b8e275604ae48485aa97403..fa76d1ff51c5d85814a1e3f3c57d80637bd0a8fc 100644
--- a/build.sh
+++ b/build.sh
@@ -23,7 +23,7 @@ build_frontend_fake() {
 
 build_backend() {
     cd $BUILD_DIR
-    cmake ..
+    cmake .. ${PYTHON_FLAGS}
     make -j2
 }
 
diff --git a/demo/mxnet/TUTORIAL_CN.md b/demo/mxnet/TUTORIAL_CN.md
index 2ac270c6b625e714026e4e4ec77f3b13687b13f1..ae1e8bd895a2e83ae5f3346a152591cba614bab0 100644
--- a/demo/mxnet/TUTORIAL_CN.md
+++ b/demo/mxnet/TUTORIAL_CN.md
@@ -24,7 +24,7 @@ VisualDL的安装很简单。请按照VisualDL的[官方网站](https://github.c
 
 ```
 python setup.py bdist_wheel
-pip install --upgrade dist/visualdl-0.0.1-py2-none-any.whl
+pip install --upgrade dist/visualdl-*.whl
 ```
 
 ## 开始编写训练MNIST的程序
diff --git a/setup.py b/setup.py
index 599c5dca50526418aee447a11c0f98bd273fde4a..23bbf0092c502a6ff7096327a2db063b36f1ee77 100644
--- a/setup.py
+++ b/setup.py
@@ -6,7 +6,7 @@ from distutils.spawn import find_executable
 from distutils import sysconfig, dep_util, log
 import setuptools.command.build_py
 import setuptools
-from setuptools import setup, find_packages
+from setuptools import setup, find_packages, Distribution, Extension
 import subprocess
 
 TOP_DIR = os.path.realpath(os.path.dirname(__file__))
@@ -79,22 +79,25 @@ datas = []
 data_root = os.path.join(TOP_DIR, 'visualdl/server/dist')
 for root, dirs, files in os.walk(data_root):
     for filename in files:
-        path = 'dist/'+os.path.join(root, filename)[len(data_root)+1:]
+        path = 'dist/' + os.path.join(root, filename)[len(data_root) + 1:]
         datas.append(path)
 print datas
 
 setup(
     name="visualdl",
     version=VERSION_NUMBER,
-    author="PaddlePaddle and Echarts team.",
-    description="Visualize Deep Learning.",
+    author="PaddlePaddle and Echarts team",
+    description="Visualize Deep Learning",
     license=LICENSE,
     keywords="visualization deeplearning",
     long_description=read('README.md'),
     install_requires=install_requires,
-    package_data={'visualdl.server': datas,
-                  'visualdl':['core.so'],
-                  'visualdl.python':['core.so', 'dog.jpg']},
+    package_data={
+        'visualdl.server': datas,
+        'visualdl': ['core.so'],
+        'visualdl.python': ['core.so', 'dog.jpg']
+    },
     packages=packages,
+    ext_modules=[Extension('_foo', ['stub.cc'])],
     scripts=['visualdl/server/visualDL', 'demo/vdl_scratch.py'],
     cmdclass=cmdclass)
diff --git a/stub.cc b/stub.cc
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/tests.sh b/tests.sh
index 12458d3fcf804376c30b5988fb40e205a3b30b2b..ace3f66e260012076096c3aeb531b50463807539 100644
--- a/tests.sh
+++ b/tests.sh
@@ -6,12 +6,13 @@ readonly TOP_DIR=$(pwd)
 readonly core_path=$TOP_DIR/build/visualdl/logic
 readonly python_path=$TOP_DIR/visualdl/python
 readonly max_file_size=1000000 # 1MB
+readonly version_number=`cat VERSION_NUMBER`
 
 sudo="sudo"
 
 if [[ "$TRAVIS_OS_NAME" == "osx" ]]; then sudo=""; fi
 
-if [[ "$TRAVIS_OS_NAME" == "osx" ]]; then 
+if [[ "$TRAVIS_OS_NAME" == "osx" ]]; then
 		curl -O http://python-distribute.org/distribute_setup.py
 		python distribute_setup.py
 		curl -O https://raw.github.com/pypa/pip/master/contrib/get-pip.py
@@ -41,7 +42,7 @@ package() {
 
     cd $TOP_DIR
     python setup.py bdist_wheel
-    $sudo pip install dist/visualdl-0.0.1-py2-none-any.whl
+    $sudo pip install dist/visualdl-${version_number}-*.whl
 }
 
 backend_test() {
diff --git a/visualdl/logic/CMakeLists.txt b/visualdl/logic/CMakeLists.txt
index 415bd93438e31fe0938e7c4fe6da2fbac87d5b04..acee65cb97b49ad7e2d1c928ca6f3b7683cecbb0 100644
--- a/visualdl/logic/CMakeLists.txt
+++ b/visualdl/logic/CMakeLists.txt
@@ -18,12 +18,19 @@ add_dependencies(im storage_proto)
 add_dependencies(sdk entry storage storage_proto eigen3)
 
 ## pybind
+set(OPTIONAL_LINK_FLAGS)
+
+if(NOT APPLE)
+  set(OPTIONAL_LINK_FLAGS "rt")
+endif()
+
 add_library(core SHARED ${PROJECT_SOURCE_DIR}/visualdl/logic/pybind.cc)
+
 if (NOT ON_RELEASE)
   add_dependencies(core pybind python im entry tablet storage sdk protobuf glog eigen3)
-  target_link_libraries(core PRIVATE pybind entry python im tablet storage sdk protobuf glog)
+  target_link_libraries(core PRIVATE pybind entry python im tablet storage sdk protobuf glog ${OPTIONAL_LINK_FLAGS})
 else()
   add_dependencies(core pybind python im entry tablet storage sdk protobuf eigen3)
-  target_link_libraries(core PRIVATE pybind entry python im tablet storage sdk protobuf)
+  target_link_libraries(core PRIVATE pybind entry python im tablet storage sdk protobuf ${OPTIONAL_LINK_FLAGS})
 endif()
 set_target_properties(core PROPERTIES PREFIX "" SUFFIX ".so")