diff --git a/python/CMakeLists.txt b/python/CMakeLists.txt
index 9cdcb87df5dd1669066c204c86c269973df506f1..1c5ded943b3814688af1f177503d3bdc35073c3f 100644
--- a/python/CMakeLists.txt
+++ b/python/CMakeLists.txt
@@ -87,6 +87,7 @@ if (WITH_TESTING)
     endif()
   endif()
   add_subdirectory(paddle/fluid/tests)
+  add_subdirectory(paddle/fluid/contrib/tests)
 endif()
 install(DIRECTORY ${PADDLE_PYTHON_PACKAGE_DIR}
     DESTINATION opt/paddle/share/wheels
diff --git a/python/paddle/fluid/__init__.py b/python/paddle/fluid/__init__.py
index c2cf7dd84ea149d345ac7a148de842d7e3965c24..9aac3c7fc16ae1ded2700662764895385b043130 100644
--- a/python/paddle/fluid/__init__.py
+++ b/python/paddle/fluid/__init__.py
@@ -47,7 +47,7 @@ from .param_attr import ParamAttr, WeightNormParamAttr
 from .data_feeder import DataFeeder
 from .core import LoDTensor, LoDTensorArray, CPUPlace, CUDAPlace, CUDAPinnedPlace, Scope
 from .transpiler import DistributeTranspiler, InferenceTranspiler, \
-    memory_optimize, release_memory, DistributeTranspilerConfig, QuantizeTranspiler
+    memory_optimize, release_memory, DistributeTranspilerConfig
 from .lod_tensor import create_lod_tensor, create_random_int_lodtensor
 from . import clip
 from . import profiler
diff --git a/python/paddle/fluid/contrib/__init__.py b/python/paddle/fluid/contrib/__init__.py
index 5607f11932bbe6aff548be316dc39b4636e079f4..2d78af05ecc05ac06a6c5caf55c0697947ebea57 100644
--- a/python/paddle/fluid/contrib/__init__.py
+++ b/python/paddle/fluid/contrib/__init__.py
@@ -18,5 +18,7 @@ from . import decoder
 from .decoder import *
 from . import memory_usage_calc
 from .memory_usage_calc import *
+from . import quantize
 
 __all__ = decoder.__all__ + memory_usage_calc.__all__
+__all__ += quantize.__all__
diff --git a/python/paddle/fluid/contrib/quantize/__init__.py b/python/paddle/fluid/contrib/quantize/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..4e1f3d8dc40ba796f8b07b82ae6bb1eee723da33
--- /dev/null
+++ b/python/paddle/fluid/contrib/quantize/__init__.py
@@ -0,0 +1,19 @@
+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+# 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.
+
+from __future__ import print_function
+
+from . import quantize_transpiler
+
+__all__ = quantize_transpiler.__all__
diff --git a/python/paddle/fluid/transpiler/quantize_transpiler.py b/python/paddle/fluid/contrib/quantize/quantize_transpiler.py
similarity index 98%
rename from python/paddle/fluid/transpiler/quantize_transpiler.py
rename to python/paddle/fluid/contrib/quantize/quantize_transpiler.py
index ec8193760bdb95eebb6891d1ff1791e139091511..b04c35082748f88bf33cb4d5fa2f275eb8aefb5e 100644
--- a/python/paddle/fluid/transpiler/quantize_transpiler.py
+++ b/python/paddle/fluid/contrib/quantize/quantize_transpiler.py
@@ -18,14 +18,16 @@ import numpy as np
 from paddle.fluid.framework import default_main_program, default_startup_program, program_guard
 from paddle.fluid.layer_helper import LayerHelper
 from paddle.fluid import unique_name
+from paddle.fluid import core
 from paddle.fluid.initializer import Constant
 from paddle.fluid.param_attr import ParamAttr
 from paddle.fluid.layer_helper import LayerHelper
 from paddle.fluid.layers.nn import autoincreased_step_counter
-from .. import core
-from ..framework import Variable
-from ..executor import global_scope
-from inference_transpiler import InferenceTranspiler
+from paddle.fluid.framework import Variable
+from paddle.fluid.executor import global_scope
+from paddle.fluid.transpiler.inference_transpiler import InferenceTranspiler
+
+__all__ = ['QuantizeTranspiler']
 
 _QUANTIZABLE_OP_TYPES = ['conv2d', 'depthwise_conv2d', 'mul']
 
diff --git a/python/paddle/fluid/tests/transpiler/CMakeLists.txt b/python/paddle/fluid/contrib/tests/CMakeLists.txt
similarity index 100%
rename from python/paddle/fluid/tests/transpiler/CMakeLists.txt
rename to python/paddle/fluid/contrib/tests/CMakeLists.txt
diff --git a/python/paddle/fluid/tests/transpiler/test_quantize_transpiler.py b/python/paddle/fluid/contrib/tests/test_quantize_transpiler.py
similarity index 94%
rename from python/paddle/fluid/tests/transpiler/test_quantize_transpiler.py
rename to python/paddle/fluid/contrib/tests/test_quantize_transpiler.py
index 5245b5ea09918d830fcdeb541c65810b139a790e..d4e161ad8ae936496bea5015b7bd03f2d55d5249 100644
--- a/python/paddle/fluid/tests/transpiler/test_quantize_transpiler.py
+++ b/python/paddle/fluid/contrib/tests/test_quantize_transpiler.py
@@ -13,17 +13,20 @@
 # limitations under the license.
 
 import numpy as np
+import six
+
 import unittest
 import paddle
 import paddle.fluid as fluid
-from paddle.fluid.transpiler.quantize_transpiler import _original_var_name
+from paddle.fluid.contrib.quantize.quantize_transpiler import _original_var_name
+from paddle.fluid.contrib.quantize.quantize_transpiler import QuantizeTranspiler
 
 
 def linear_fc(num):
     data = fluid.layers.data(name='image', shape=[1, 32, 32], dtype='float32')
     label = fluid.layers.data(name='label', shape=[1], dtype='int64')
     hidden = data
-    for _ in xrange(num):
+    for _ in six.moves.xrange(num):
         hidden = fluid.layers.fc(hidden, size=128, act='relu')
     loss = fluid.layers.cross_entropy(input=hidden, label=label)
     loss = fluid.layers.mean(loss)
@@ -51,7 +54,7 @@ def residual_block(num):
     data = fluid.layers.data(name='image', shape=[1, 32, 32], dtype='float32')
     label = fluid.layers.data(name='label', shape=[1], dtype='int64')
     hidden = data
-    for _ in xrange(num):
+    for _ in six.moves.xrange(num):
         conv = conv_bn_layer(hidden, 16, 3, 1, 1, act=None, bias_attr=True)
         short = conv_bn_layer(hidden, 16, 1, 1, 0, act=None)
         hidden = fluid.layers.elementwise_add(x=conv, y=short, act='relu')
@@ -142,7 +145,7 @@ class TestQuantizeTranspiler(unittest.TestCase):
             loss = linear_fc(3)
             opt = fluid.optimizer.Adam(learning_rate=0.001)
             opt.minimize(loss)
-            t = fluid.QuantizeTranspiler(activation_quantize_type=quant_type)
+            t = QuantizeTranspiler(activation_quantize_type=quant_type)
             t.training_transpile(main)
             self.check_program(main)
 
@@ -161,7 +164,7 @@ class TestQuantizeTranspiler(unittest.TestCase):
             loss = residual_block(2)
             opt = fluid.optimizer.Adam(learning_rate=0.001)
             opt.minimize(loss)
-            t = fluid.QuantizeTranspiler(activation_quantize_type=quant_type)
+            t = QuantizeTranspiler(activation_quantize_type=quant_type)
             t.training_transpile(main)
             self.check_program(main)
 
@@ -176,7 +179,7 @@ class TestQuantizeTranspiler(unittest.TestCase):
     def freeze_program(self, use_cuda):
         main = fluid.Program()
         startup = fluid.Program()
-        quant_transpiler = fluid.QuantizeTranspiler()
+        quant_transpiler = QuantizeTranspiler()
         with fluid.program_guard(main, startup):
             img = fluid.layers.data(
                 name='image', shape=[1, 28, 28], dtype='float32')
@@ -247,7 +250,11 @@ class TestQuantizeTranspiler(unittest.TestCase):
         self.assertEqual(np.sum(w_8bit), np.sum(w_freeze))
 
     def test_freeze_program_cuda(self):
-        self.freeze_program(True)
+        if fluid.core.is_compiled_with_cuda():
+            self.freeze_program(True)
+
+    def test_freeze_program_cpu(self):
+        self.freeze_program(False)
 
 
 if __name__ == '__main__':
diff --git a/python/paddle/fluid/tests/CMakeLists.txt b/python/paddle/fluid/tests/CMakeLists.txt
index 32447dc11e2f37c18b267a7d13b49718854bc56c..d24417bbacb503d9ea70e68e7e0edb59e7dddbde 100644
--- a/python/paddle/fluid/tests/CMakeLists.txt
+++ b/python/paddle/fluid/tests/CMakeLists.txt
@@ -8,4 +8,3 @@ endforeach()
 add_subdirectory(unittests)
 add_subdirectory(book)
 add_subdirectory(book_memory_optimization)
-add_subdirectory(transpiler)
diff --git a/python/paddle/fluid/transpiler/__init__.py b/python/paddle/fluid/transpiler/__init__.py
index 16199890d66731f4c77cfa82601cb0580310d62e..79557a2cc04ff5cb4c0f974c98c05d512cc2d53f 100644
--- a/python/paddle/fluid/transpiler/__init__.py
+++ b/python/paddle/fluid/transpiler/__init__.py
@@ -16,12 +16,15 @@ from __future__ import print_function
 
 from .distribute_transpiler import DistributeTranspiler, DistributeTranspilerConfig
 from .inference_transpiler import InferenceTranspiler
-from .quantize_transpiler import QuantizeTranspiler
 from .memory_optimization_transpiler import memory_optimize, release_memory
 from .ps_dispatcher import HashName, RoundRobin
 
 __all__ = [
-    "DistributeTranspiler", "InferenceTranspiler", "memory_optimize",
-    "release_memory", "HashName", "RoundRobin", "DistributeTranspilerConfig",
-    "QuantizeTranspiler"
+    "DistributeTranspiler",
+    "InferenceTranspiler",
+    "memory_optimize",
+    "release_memory",
+    "HashName",
+    "RoundRobin",
+    "DistributeTranspilerConfig",
 ]