Added test that demonstrates that keras custom layers are not supported for clustering.

Change-Id: I235fb95986cc9219f6afb6deb95e6d7e47631476

Author: wwwind

tensorflow_model_optimization/python/core/clustering/keras/cluster_test.py

@@ -53,6 +53,25 @@ def get_clusterable_weights(self):
 class CustomNonClusterableLayer(layers.Dense):
   pass
 
+class KerasCustomLayer(keras.layers.Layer):
+  def __init__(self, units=32):
+    super(KerasCustomLayer, self).__init__()
+    self.units = units
+
+  def build(self, input_shape):
+    self.w = self.add_weight(
+      shape=(input_shape[-1], self.units),
+      initializer="random_normal",
+      trainable=True,
+    )
+    self.b = self.add_weight(
+      shape=(self.units,),
+      initializer="random_normal",
+      trainable=False
+    )
+
+  def call(self, inputs):
+    return tf.matmul(inputs, self.w) + self.b
 
 class ClusterTest(test.TestCase, parameterized.TestCase):
   """Unit tests for the cluster module."""
@@ -66,6 +85,7 @@ def setUp(self):
     self.custom_clusterable_layer = CustomClusterableLayer(10)
     self.custom_non_clusterable_layer = CustomNonClusterableLayer(10)
     self.keras_depthwiseconv2d_layer = layers.DepthwiseConv2D((3, 3), (1, 1))
+    self.keras_custom_layer = KerasCustomLayer()
 
     clustering_registry.ClusteringLookupRegistry.register_new_implementation(
         {
@@ -179,6 +199,22 @@ def testClusterCustomNonClusterableLayer(self):
       cluster_wrapper.ClusterWeights(custom_non_clusterable_layer,
                                      **self.params)
 
+  def testClusterKerasCustomLayer(self):
+    """
+    Verifies that attempting to cluster a keras custom layer raises
+    an exception.
+    """
+    # If layer is not built, it has not weights, so
+    # we just skip it.
+    keras_custom_layer = self.keras_custom_layer
+    cluster_wrapper.ClusterWeights(keras_custom_layer,
+                                  **self.params)
+    # We need to build weights before check that clustering is not supported.
+    keras_custom_layer.build(input_shape=(10, 10))
+    with self.assertRaises(ValueError):
+      cluster_wrapper.ClusterWeights(keras_custom_layer,
+                                     **self.params)
+
   @keras_parameterized.run_all_keras_modes
   def testClusterSequentialModelSelectively(self):
     """