diff --git a/python/paddle/fluid/tests/unittests/ir/inference/inference_pass_test.py b/python/paddle/fluid/tests/unittests/ir/inference/inference_pass_test.py
index 95b812bdf7ca48562b66f50e55ae443780857c72..b2a27f958868d2dcd97c645c051618e9a46c7283 100644
--- a/python/paddle/fluid/tests/unittests/ir/inference/inference_pass_test.py
+++ b/python/paddle/fluid/tests/unittests/ir/inference/inference_pass_test.py
@@ -53,6 +53,9 @@ class InferencePassTest(unittest.TestCase):
                             feed=self.feeds,
                             fetch_list=self.fetch_list,
                             return_numpy=False)
+        # save models as combined to ensure that 
+        # there won't be too many useless files 
+        # after finishing a couple of tests.
         fluid.io.save_inference_model(
             dirname=self.path,
             feeded_var_names=list(self.feeds.keys()),
diff --git a/python/paddle/fluid/tests/unittests/ir/inference/test_transpose_flatten_concat_fuse_pass.py b/python/paddle/fluid/tests/unittests/ir/inference/test_transpose_flatten_concat_fuse_pass.py
new file mode 100644
index 0000000000000000000000000000000000000000..5c91ceeb44a0352ea9a753fe7475496048042b1f
--- /dev/null
+++ b/python/paddle/fluid/tests/unittests/ir/inference/test_transpose_flatten_concat_fuse_pass.py
@@ -0,0 +1,51 @@
+# Copyright (c) 2020 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.
+
+import unittest
+import numpy as np
+from inference_pass_test import InferencePassTest
+import paddle.fluid as fluid
+import paddle.fluid.core as core
+
+
+class TransposeFlattenConcatFusePassTest(InferencePassTest):
+    def setUp(self):
+        with fluid.program_guard(self.main_program, self.startup_program):
+            data1 = fluid.data(
+                name="data1", shape=[8, 32, 128], dtype="float32")
+            data2 = fluid.data(
+                name="data2", shape=[8, 32, 128], dtype="float32")
+            trans1 = fluid.layers.transpose(data1, perm=[2, 1, 0])
+            trans2 = fluid.layers.transpose(data2, perm=[2, 1, 0])
+            flatt1 = fluid.layers.flatten(trans1)
+            flatt2 = fluid.layers.flatten(trans2)
+            concat_out = fluid.layers.concat([flatt1, flatt2])
+            # There is no parameters for above structure. 
+            # Hence, append a batch_norm to avoid failure caused by load_combined. 
+            out = fluid.layers.batch_norm(concat_out, is_test=True)
+
+        self.feeds = {
+            "data1": np.random.random([8, 32, 128]).astype("float32"),
+            "data2": np.random.random([8, 32, 128]).astype("float32")
+        }
+        self.fetch_list = [out]
+
+    def test_check_output(self):
+        # There is no cpu pass for transpose_flatten_concat_fuse
+        if core.is_compiled_with_cuda():
+            self.check_output_with_option([True])
+
+
+if __name__ == "__main__":
+    unittest.main()
diff --git a/python/paddle/fluid/tests/unittests/ir/inference/test_trt_transpose_flatten_concat_fuse_pass.py b/python/paddle/fluid/tests/unittests/ir/inference/test_trt_transpose_flatten_concat_fuse_pass.py
new file mode 100644
index 0000000000000000000000000000000000000000..c85c54c74187f0ca77cd62928d15040a4b70adde
--- /dev/null
+++ b/python/paddle/fluid/tests/unittests/ir/inference/test_trt_transpose_flatten_concat_fuse_pass.py
@@ -0,0 +1,55 @@
+# Copyright (c) 2020 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.
+
+import unittest
+import numpy as np
+from inference_pass_test import InferencePassTest
+import paddle.fluid as fluid
+import paddle.fluid.core as core
+from paddle.fluid.core import AnalysisConfig
+
+
+class TransposeFlattenConcatFusePassTRTTest(InferencePassTest):
+    def setUp(self):
+        with fluid.program_guard(self.main_program, self.startup_program):
+            data1 = fluid.data(
+                name="data1", shape=[8, 32, 128], dtype="float32")
+            data2 = fluid.data(
+                name="data2", shape=[8, 32, 128], dtype="float32")
+            trans1 = fluid.layers.transpose(data1, perm=[2, 1, 0])
+            trans2 = fluid.layers.transpose(data2, perm=[2, 1, 0])
+            flatt1 = fluid.layers.flatten(trans1)
+            flatt2 = fluid.layers.flatten(trans2)
+            concat_out = fluid.layers.concat([flatt1, flatt2])
+            # There is no parameters for above structure. 
+            # Hence, append a batch_norm to avoid failure caused by load_combined. 
+            out = fluid.layers.batch_norm(concat_out, is_test=True)
+
+        self.feeds = {
+            "data1": np.random.random([8, 32, 128]).astype("float32"),
+            "data2": np.random.random([8, 32, 128]).astype("float32")
+        }
+        self.enable_trt = True
+        self.trt_parameters = TransposeFlattenConcatFusePassTRTTest.TensorRTParam(
+            1 << 20, 1, 3, AnalysisConfig.Precision.Float32, False, False)
+        self.fetch_list = [out]
+
+    def test_check_output(self):
+        # There is no cpu pass for transpose_flatten_concat_fuse
+        if core.is_compiled_with_cuda():
+            self.check_output_with_option([True])
+
+
+if __name__ == "__main__":
+    unittest.main()
diff --git a/python/paddle/fluid/tests/unittests/test_inference_api.py b/python/paddle/fluid/tests/unittests/test_inference_api.py
index a74771211f6d3cd432c22afb7c80ec5a538a168a..98ec0b3db04c496ead3c6af409264031953dbf37 100644
--- a/python/paddle/fluid/tests/unittests/test_inference_api.py
+++ b/python/paddle/fluid/tests/unittests/test_inference_api.py
@@ -24,50 +24,34 @@ class TestInferenceApi(unittest.TestCase):
     def test_inference_api(self):
         tensor32 = np.random.randint(10, 20, size=[20, 2]).astype('int32')
         paddletensor32 = PaddleTensor(tensor32)
-        value32 = np.array(paddletensor32.data.int32_data()).reshape(* [20, 2])
         dtype32 = paddletensor32.dtype
-        self.assertEqual(value32.all(), tensor32.all())
         self.assertEqual(dtype32, PaddleDType.INT32)
-        self.assertEqual(
-            type(paddletensor32.data.tolist('int32')), type(tensor32.tolist()))
         self.assertEqual(
             paddletensor32.data.tolist('int32'), tensor32.ravel().tolist())
-        self.assertEqual(type(paddletensor32.as_ndarray()), type(tensor32))
         paddletensor32.data.reset(tensor32)
-        self.assertEqual(paddletensor32.as_ndarray().all(), tensor32.all())
+        self.assertEqual(paddletensor32.as_ndarray().ravel().tolist(),
+                         tensor32.ravel().tolist())
 
         tensor64 = np.random.randint(10, 20, size=[20, 2]).astype('int64')
         paddletensor64 = PaddleTensor(tensor64)
-        value64 = np.array(paddletensor64.data.int64_data()).reshape(* [20, 2])
         dtype64 = paddletensor64.dtype
-        self.assertEqual(value64.all(), tensor64.all())
         self.assertEqual(dtype64, PaddleDType.INT64)
-        self.assertEqual(
-            type(paddletensor64.data.tolist('int64')), type(tensor64.tolist()))
         self.assertEqual(
             paddletensor64.data.tolist('int64'), tensor64.ravel().tolist())
-        self.assertEqual(type(paddletensor64.as_ndarray()), type(tensor64))
         paddletensor64.data.reset(tensor64)
-        self.assertEqual(paddletensor64.as_ndarray().all(), tensor64.all())
+        self.assertEqual(paddletensor64.as_ndarray().ravel().tolist(),
+                         tensor64.ravel().tolist())
 
         tensor_float = np.random.randn(20, 2).astype('float32')
         paddletensor_float = PaddleTensor(tensor_float)
-        value_float = np.array(paddletensor_float.data.float_data()).reshape(
-            * [20, 2])
         dtype_float = paddletensor_float.dtype
-        self.assertEqual(value_float.all(), tensor_float.all())
         self.assertEqual(dtype_float, PaddleDType.FLOAT32)
-        self.assertEqual(
-            type(paddletensor_float.data.tolist('float32')),
-            type(tensor_float.tolist()))
         self.assertEqual(
             paddletensor_float.data.tolist('float32'),
             tensor_float.ravel().tolist())
-        self.assertEqual(
-            type(paddletensor_float.as_ndarray()), type(tensor_float))
         paddletensor_float.data.reset(tensor_float)
-        self.assertEqual(paddletensor_float.as_ndarray().all(),
-                         tensor_float.all())
+        self.assertEqual(paddletensor_float.as_ndarray().ravel().tolist(),
+                         tensor_float.ravel().tolist())
 
 
 if __name__ == '__main__':