convert test_session to cached_session to avoid deprecation warning

deepcell/layers/filter_detections_test.py

@@ -40,7 +40,7 @@ class TestFilterDetections(test.TestCase):
 
     @tf_test_util.run_in_graph_and_eager_modes()
     def test_simple(self):
-        with self.test_session(use_gpu=True):
+        with self.cached_session(use_gpu=True):
             # create simple FilterDetections layer
             layer = layers.FilterDetections()
 
@@ -81,7 +81,7 @@ def test_simple(self):
 
     @tf_test_util.run_in_graph_and_eager_modes()
     def test_simple_with_other(self):
-        with self.test_session(use_gpu=True):
+        with self.cached_session(use_gpu=True):
             # create simple FilterDetections layer
             layer = layers.FilterDetections()
 
@@ -142,7 +142,7 @@ def test_simple_with_other(self):
 
     # @tf_test_util.run_in_graph_and_eager_modes()
     # def test_mini_batch(self):
-    #     with self.test_session(use_gpu=True):
+    #     with self.cached_session(use_gpu=True):
     #         # create simple FilterDetections layer
     #         layer = layers.FilterDetections()
     #

deepcell/layers/location_test.py

@@ -39,7 +39,7 @@ class LocationTest(test.TestCase):
 
     @tf_test_util.run_in_graph_and_eager_modes()
     def test_location_2d(self):
-        with self.test_session(use_gpu=True):
+        with self.cached_session(use_gpu=True):
             testing_utils.layer_test(
                 layers.Location2D,
                 kwargs={'in_shape': (5, 6, 4),
@@ -55,7 +55,7 @@ def test_location_2d(self):
 
     @tf_test_util.run_in_graph_and_eager_modes()
     def test_location_3d(self):
-        with self.test_session(use_gpu=True):
+        with self.cached_session(use_gpu=True):
             testing_utils.layer_test(
                 layers.Location3D,
                 kwargs={'in_shape': (11, 12, 10, 4),

deepcell/layers/normalization_test.py

@@ -43,7 +43,7 @@ class ImageNormalizationTest(test.TestCase):
     def test_normalize_2d(self):
         custom_objects = {'ImageNormalization2D': layers.ImageNormalization2D}
         norm_methods = [None, 'std', 'max', 'whole_image']
-        with self.test_session(use_gpu=True):
+        with self.cached_session(use_gpu=True):
             # test each norm method
             for norm_method in norm_methods:
                 testing_utils.layer_test(
@@ -86,7 +86,7 @@ def test_normalize_2d(self):
     def test_normalize_3d(self):
         custom_objects = {'ImageNormalization3D': layers.ImageNormalization3D}
         norm_methods = [None, 'std', 'max', 'whole_image']
-        with self.test_session(use_gpu=True):
+        with self.cached_session(use_gpu=True):
             # test each norm method
             for norm_method in norm_methods:
                 testing_utils.layer_test(

deepcell/layers/padding_test.py

@@ -60,7 +60,7 @@ def test_reflection_padding_2d(self):
         data_formats = ['channels_first', 'channels_last']
         for data_format, inputs in zip(data_formats, [ins2, ins1]):
             # basic test
-            with self.test_session(use_gpu=True):
+            with self.cached_session(use_gpu=True):
                 testing_utils.layer_test(
                     layers.ReflectionPadding2D,
                     kwargs={'padding': (2, 2),
@@ -75,7 +75,7 @@ def test_reflection_padding_2d(self):
                     input_shape=inputs.shape)
 
             # correctness test
-            # with self.test_session(use_gpu=True):
+            # with self.cached_session(use_gpu=True):
             #     layer = layers.ReflectionPadding2D(
             #         padding=(2, 2), data_format=data_format)
             #     layer.build(inputs.shape)
@@ -146,7 +146,7 @@ def test_reflection_padding_3d(self):
         data_formats = ['channels_first', 'channels_last']
         for data_format, inputs in zip(data_formats, [inputs2, inputs1]):
             # basic test
-            with self.test_session(use_gpu=True):
+            with self.cached_session(use_gpu=True):
                 testing_utils.layer_test(
                     layers.ReflectionPadding3D,
                     kwargs={'padding': (2, 2, 2),
@@ -155,7 +155,7 @@ def test_reflection_padding_3d(self):
                     input_shape=inputs.shape)
 
         # correctness test
-        # with self.test_session(use_gpu=True):
+        # with self.cached_session(use_gpu=True):
         #     layer = layers.ReflectionPadding3D(padding=(2, 2, 2))
         #     layer.build(inputs.shape)
         #     output = layer(keras.backend.variable(inputs))

deepcell/layers/pooling_test.py

@@ -44,7 +44,7 @@ def test_dilated_max_pool_2d(self):
         for strides in [(1, 1), (2, 2), None]:
             for dilation_rate in [1, 2]:
                 for padding in ['valid', 'same']:
-                    with self.test_session(use_gpu=True):
+                    with self.cached_session(use_gpu=True):
                         testing_utils.layer_test(
                             layers.DilatedMaxPool2D,
                             kwargs={'strides': strides,
@@ -71,7 +71,7 @@ def test_dilated_max_pool_3d(self):
         for strides in [1, 2, None]:
             for dilation_rate in [1, 2]:
                 for padding in ['valid', 'same']:
-                    with self.test_session(use_gpu=True):
+                    with self.cached_session(use_gpu=True):
                         testing_utils.layer_test(
                             layers.DilatedMaxPool3D,
                             kwargs={'strides': strides,

deepcell/layers/resize_test.py

@@ -39,7 +39,7 @@ class ResizeTest(test.TestCase):
 
     @tf_test_util.run_in_graph_and_eager_modes()
     def test_resize_2d(self):
-        with self.test_session(use_gpu=True):
+        with self.cached_session(use_gpu=True):
             testing_utils.layer_test(
                 layers.Resize2D,
                 kwargs={'scale': 2},

deepcell/layers/retinanet_test.py

@@ -41,7 +41,7 @@ class TestAnchors(test.TestCase):
 
     @tf_test_util.run_in_graph_and_eager_modes()
     def test_anchors_2d(self):
-        with self.test_session(use_gpu=True):
+        with self.cached_session(use_gpu=True):
             testing_utils.layer_test(
                 layers.Anchors,
                 kwargs={'size': 1, 'stride': 1,
@@ -63,7 +63,7 @@ def test_anchors_2d(self):
 
     @tf_test_util.run_in_graph_and_eager_modes()
     def test_simple(self):
-        with self.test_session():
+        with self.cached_session():
             # create simple Anchors layer
             anchors_layer = layers.Anchors(
                 size=32,
@@ -93,7 +93,7 @@ def test_simple(self):
 
     @tf_test_util.run_in_graph_and_eager_modes()
     def test_mini_batch(self):
-        with self.test_session():
+        with self.cached_session():
             # create simple Anchors layer
             anchors_layer = layers.Anchors(
                 size=32,
@@ -127,7 +127,7 @@ class TestRegressBoxes(test.TestCase):
 
     @tf_test_util.run_in_graph_and_eager_modes()
     def test_simple(self):
-        with self.test_session():
+        with self.cached_session():
             # create simple RegressBoxes layer
             layer = layers.RegressBoxes()
 
@@ -165,7 +165,7 @@ def test_simple(self):
 
     @tf_test_util.run_in_graph_and_eager_modes()
     def test_mini_batch(self):
-        with self.test_session():
+        with self.cached_session():
             mean = [0, 0, 0, 0]
             std = [0.2, 0.2, 0.2, 0.2]
 
@@ -257,7 +257,7 @@ def test_simple(self):
         ]], dtype=K.floatx())
 
         # test channels_last
-        with self.test_session():
+        with self.cached_session():
             # create input
             image = K.variable(np.random.random((1, img_h, img_w, 3)))
 
@@ -274,7 +274,7 @@ def test_simple(self):
             self.assertAllClose(actual, expected)
 
         # test channels_first
-        with self.test_session():
+        with self.cached_session():
             # create input
             image = K.variable(np.random.random((1, 6, img_h, img_w)))
 

deepcell/layers/tensor_product_test.py

@@ -92,7 +92,7 @@ def test_tensorproduct(self):
                 input_shape=(3, 5, 6, None))
 
     def test_tensorproduct_regularization(self):
-        with self.test_session():
+        with self.cached_session():
             layer = layers.TensorProduct(
                 3,
                 kernel_regularizer=keras.regularizers.l1(0.01),
@@ -103,7 +103,7 @@ def test_tensorproduct_regularization(self):
             self.assertEqual(3, len(layer.losses))
 
     def test_tensorproduct_constraints(self):
-        with self.test_session():
+        with self.cached_session():
             k_constraint = keras.constraints.max_norm(0.01)
             b_constraint = keras.constraints.max_norm(0.01)
             layer = layers.TensorProduct(

deepcell/layers/upsample_test.py

@@ -42,7 +42,7 @@ class TestUpsampleLike(test.TestCase):
     @tf_test_util.run_in_graph_and_eager_modes()
     def test_simple(self):
         # channels_last
-        with self.test_session():
+        with self.cached_session():
             # create simple UpsampleLike layer
             upsample_like_layer = layers.UpsampleLike()
 
@@ -63,7 +63,7 @@ def test_simple(self):
             self.assertEqual(actual.shape, computed_shape)
             self.assertAllEqual(actual, expected)
         # channels_first
-        with self.test_session():
+        with self.cached_session():
             # create simple UpsampleLike layer
             upsample_like_layer = layers.UpsampleLike(
                 data_format='channels_first')
@@ -86,7 +86,7 @@ def test_simple(self):
 
     @tf_test_util.run_in_graph_and_eager_modes()
     def test_mini_batch(self):
-        with self.test_session():
+        with self.cached_session():
             # create simple UpsampleLike layer
             upsample_like_layer = layers.UpsampleLike()
 

deepcell/model_zoo/featurenet.py

@@ -229,8 +229,19 @@ def bn_feature_net_skip_2D(receptive_field=61,
         else:
             model_input = img
 
-        new_input_shape = model_input.get_shape().as_list()[1:]
-        models.append(bn_feature_net_2D(receptive_field=receptive_field, input_shape=new_input_shape, norm_method=None, dilated=True, padding=True, padding_mode=padding_mode, **kwargs))
+        _model_kwargs = {
+            'receptive_field': receptive_field,
+            'input_shape': model_input.get_shape().as_list()[1:],
+            'norm_method': None,
+            'dilated': True,
+            'padding': True,
+            'padding_mode': padding_mode,
+        }
+        for k in kwargs:
+            if k not in _model_kwargs:
+                _model_kwargs[k] = kwargs[k]
+
+        models.append(bn_feature_net_2D(**_model_kwargs))
         model_outputs.append(models[-1](model_input))
 
     if last_only: