diff --git a/python/mxnet/gluon/block.py b/python/mxnet/gluon/block.py
index bd6ee6a2cdcf..5ab95f65eccf 100644
--- a/python/mxnet/gluon/block.py
+++ b/python/mxnet/gluon/block.py
@@ -962,7 +962,7 @@ def _build_cache(self, *args):
                 if name in data_names.keys():
                     arg_array.append(args[data_names[name]])
                 else:
-                    if ctx == None:
+                    if ctx is None:
                         ctx = params.get(name)._ctx_list[0]
                     arg_array.append(ndarray.random.uniform(shape=params.get(name)._shape))
             # Partition the graph.
@@ -1053,9 +1053,12 @@ def register_child(self, block, name=None):
         super(HybridBlock, self).register_child(block, name)
         self._clear_cached_op()
 
-    def hybridize(self, active=True, backend=None, backend_args={}, **kwargs):
+    def hybridize(self, active=True, backend=None, backend_args=None, **kwargs):
         self._backend = backend
-        self._backend_args = backend_args
+        if backend_args is None:
+            self._backend_args = {}
+        else:
+            self._backend_args = backend_args
         self._active = active
         self._flags = list(kwargs.items())
         self._clear_cached_op()
diff --git a/tests/python/unittest/test_subgraph_op.py b/tests/python/unittest/test_subgraph_op.py
index c4543b34aeae..000825f7febf 100644
--- a/tests/python/unittest/test_subgraph_op.py
+++ b/tests/python/unittest/test_subgraph_op.py
@@ -51,8 +51,9 @@ def network_structure_3():
     ret = ret1 + ret2
     ret = mx.sym.BatchNorm(ret)
     ret = mx.sym.BatchNorm(ret)
-    return (ret, ['data'], [(2, 3, 10, 10)])
-   
+    # Return the same and shape of 'data' and auxiliary states
+    return  (ret, ['data', *ret.list_auxiliary_states()], [(2,3,10,10), (3,), (3,), (3,), (3,)])
+
 def network_structure_4():
     # the last op has multiple duplicate outputs
     data = mx.sym.var('data', shape=(2, 3, 10, 10))
@@ -84,23 +85,24 @@ def network_structure_7():
     return (ret, ['data'], [(1,)])
 
 def get_graphs(): 
-    return [(network_structure_1(), ['Convolution']),
+    return [
+            (network_structure_1(), ['Convolution']),
             (network_structure_2(), ['exp', 'sin', '_Plus', 'elemwise_add', '_plus']),
             (network_structure_2(), ['exp', 'cos', '_Plus', 'elemwise_add', '_plus']),
-            # To do: fix batch norm issue for gluon tests.
-            #(network_structure_3(), ['exp', 'sin', '_Plus', 'elemwise_add', '_plus']),
-            #(network_structure_3(), ['exp', 'cos', '_Plus', 'elemwise_add', '_plus']),
-            #(network_structure_3(), ['exp', 'sin', '_Plus', 'elemwise_add', '_plus', 'BatchNorm']),
-            #(network_structure_3(), ['exp', 'cos', '_Plus', 'elemwise_add', '_plus', 'BatchNorm']),
-            #(network_structure_3(), ['exp', 'BatchNorm']),
-            #(network_structure_3(), ['BatchNorm']),
+            (network_structure_3(), ['exp', 'sin', '_Plus', 'elemwise_add', '_plus']),
+            (network_structure_3(), ['exp', 'cos', '_Plus', 'elemwise_add', '_plus']),
+            (network_structure_3(), ['exp', 'sin', '_Plus', 'elemwise_add', '_plus', 'BatchNorm']),
+            (network_structure_3(), ['exp', 'cos', '_Plus', 'elemwise_add', '_plus', 'BatchNorm']),
+            (network_structure_3(), ['exp', 'BatchNorm']),
+            (network_structure_3(), ['BatchNorm']),
             (network_structure_4(), ['exp']),
             (network_structure_5(), ['_plus', '_Plus', 'elemwise_add']),
             (network_structure_6(), []),
             (network_structure_6(), [mx.sym.sin.__name__]),
             (network_structure_6(), [mx.sym.Convolution.__name__]),
             (network_structure_6(), [mx.sym.sin.__name__, mx.sym.Convolution.__name__]),
-            (network_structure_7(), ['sin', 'elemwise_add', '_plus', '_Plus'])]
+            (network_structure_7(), ['sin', 'elemwise_add', '_plus', '_Plus'])
+            ]
 
 def check_subgraph_exe1(sym, subgraph_backend, op_names):
     """Use the partitioned sym to simple_bind an executor and compare the outputs