enable sequential test cases

Author: zxy844288792

tests/python/frontend/tensorflow2/test_sequential_models.py

@@ -110,46 +110,58 @@ def maxpool_batchnorm_model(input_shape, pool_size=(2, 2)):
 
 
 def test_tensorlist_stack_model():
-    class TensorArrayStackLayer(tf.keras.layers.Layer):
-        def __init__(self):
-            super().__init__()
-
-        def call(self, inputs):
-            inputs = tf.squeeze(inputs)
-            outputs = tf.TensorArray(
-                tf.float32, size=inputs.shape[0], infer_shape=False, element_shape=inputs.shape[1:]
-            )
-            outputs = outputs.unstack(inputs)
-
-            return outputs.stack()
+    def tensorlist_stack_model(input_shape):
+        class TensorArrayStackLayer(tf.keras.layers.Layer):
+            def __init__(self):
+                super().__init__()
+
+            def call(self, inputs):
+                inputs = tf.squeeze(inputs)
+                outputs = tf.TensorArray(
+                    tf.float32,
+                    size=inputs.shape[0],
+                    infer_shape=False,
+                    element_shape=inputs.shape[1:],
+                )
+                outputs = outputs.unstack(inputs)
+
+                return outputs.stack()
+
+        input_shape = (3, 32)
+        model = tf.keras.Sequential(
+            [tf.keras.layers.Input(shape=input_shape, batch_size=1), TensorArrayStackLayer()]
+        )
+        return model
 
-    shape = (3, 32)
-    model = tf.keras.Sequential(
-        [tf.keras.layers.Input(shape=shape, batch_size=1), TensorArrayStackLayer()]
-    )
-    return model
+    run_sequential_model(tensorlist_stack_model, input_shape=(3, 32))
 
 
 def test_tensorlist_read_model():
-    class TensorArrayReadLayer(tf.keras.layers.Layer):
-        def __init__(self):
-            super().__init__()
-
-        def call(self, inputs):
-            inputs = tf.squeeze(inputs)
-            outputs = tf.TensorArray(
-                tf.float32, size=inputs.shape[0], infer_shape=False, element_shape=inputs.shape[1:]
-            )
-            for i in range(inputs.shape[0]):
-                outputs = outputs.write(i, inputs[i, :])
-
-            return outputs.read(0)
-
-    shape = (3, 32)
-    model = tf.keras.Sequential(
-        [tf.keras.layers.Input(shape=shape, batch_size=1), TensorArrayReadLayer()]
-    )
-    return model
+    def tensorlist_read_model(input_shape):
+        class TensorArrayReadLayer(tf.keras.layers.Layer):
+            def __init__(self):
+                super().__init__()
+
+            def call(self, inputs):
+                inputs = tf.squeeze(inputs)
+                outputs = tf.TensorArray(
+                    tf.float32,
+                    size=inputs.shape[0],
+                    infer_shape=False,
+                    element_shape=inputs.shape[1:],
+                )
+                for i in range(inputs.shape[0]):
+                    outputs = outputs.write(i, inputs[i, :])
+
+                return outputs.read(0)
+
+        input_shape = (3, 32)
+        model = tf.keras.Sequential(
+            [tf.keras.layers.Input(shape=input_shape, batch_size=1), TensorArrayReadLayer()]
+        )
+        return model
+
+    run_sequential_model(tensorlist_read_model, input_shape=(3, 32))
 
 
 if __name__ == "__main__":