diff --git a/paddle/framework/init.cc b/paddle/framework/init.cc
index 682cff168d4d31e0565fc987604f97a671566fbd..1980023372572a4d058275d9af72eb61f145c1bc 100644
--- a/paddle/framework/init.cc
+++ b/paddle/framework/init.cc
@@ -75,5 +75,9 @@ bool InitDevices(const std::vector<std::string> &devices) {
   return true;
 }
 
+void InitGLOG(const std::string &prog_name) {
+  google::InitGoogleLogging(prog_name.c_str());
+}
+
 }  // namespace framework
 }  // namespace paddle
diff --git a/paddle/framework/init.h b/paddle/framework/init.h
index 33907f9eb00fb3469b53dcf8151557cc7a2d3791..9c84a03ded52632047841f95badbcf44bc9f48d1 100644
--- a/paddle/framework/init.h
+++ b/paddle/framework/init.h
@@ -22,6 +22,8 @@ namespace framework {
 
 void InitGflags(std::vector<std::string> &argv);
 
+void InitGLOG(const std::string &prog_name);
+
 bool InitDevices(const std::vector<std::string> &devices);
 
 }  // namespace framework
diff --git a/paddle/pybind/pybind.cc b/paddle/pybind/pybind.cc
index 04485ce7c1ab87f8655b0e6cbaecc36b3382f647..364db62cba6acd7ac380b5017d09f22eefa24813 100644
--- a/paddle/pybind/pybind.cc
+++ b/paddle/pybind/pybind.cc
@@ -427,6 +427,7 @@ All parameter, weight, gradient are variables in Paddle.
 
   m.def("unique_integer", UniqueIntegerGenerator);
   m.def("init_gflags", framework::InitGflags);
+  m.def("init_glog", framework::InitGLOG);
   m.def("init_devices", &framework::InitDevices);
 
   m.def("is_compile_gpu", IsCompileGPU);
diff --git a/python/paddle/v2/fluid/__init__.py b/python/paddle/v2/fluid/__init__.py
index 225b41c5043b5792abb90bbad53cbbfce9a3156e..b00892d91a027213527ca879c86b0c1b47b762e9 100644
--- a/python/paddle/v2/fluid/__init__.py
+++ b/python/paddle/v2/fluid/__init__.py
@@ -1,23 +1,23 @@
 # import all class inside framework into fluid module
-import framework
-from framework import *
-# import all class inside executor into fluid module
-import executor
-from executor import *
+from core import LoDTensor
 
-import io
+import backward
+import clip
 import evaluator
+# import all class inside executor into fluid module
+import executor
+import framework
 import initializer
+import io
 import layers
 import nets
 import optimizer
-import backward
 import regularizer
-from param_attr import ParamAttr
 from data_feeder import DataFeeder
-from core import LoDTensor, CPUPlace, CUDAPlace
 from distribute_transpiler import DistributeTranspiler
-import clip
+from executor import *
+from framework import *
+from param_attr import ParamAttr
 
 Tensor = LoDTensor
 __all__ = framework.__all__ + executor.__all__ + [
@@ -27,7 +27,7 @@ __all__ = framework.__all__ + executor.__all__ + [
 ]
 
 
-def __read_gflags_from_env__():
+def __bootstrap__():
     """
     Enable reading gflags from environment variables.
 
@@ -41,6 +41,7 @@ def __read_gflags_from_env__():
         read_env_flags.append('fraction_of_gpu_memory_to_use')
     core.init_gflags([sys.argv[0]] +
                      ["--tryfromenv=" + ",".join(read_env_flags)])
+    core.init_glog(sys.argv[0])
 
     if core.is_compile_gpu():
         core.init_devices(["CPU", "GPU:0"])
@@ -48,4 +49,4 @@ def __read_gflags_from_env__():
         core.init_devices(["CPU"])
 
 
-__read_gflags_from_env__()
+__bootstrap__()