def testRegularization(self, type, scale):
layer = tf.keras.layers.Dense(256)
layer.build([None, 128])
regularization = optim.regularization_penalty(
type, scale, weights_list=layer.trainable_variables)
self.assertEqual(0, len(regularization.shape.as_list()))
if not compat.is_tf2():
with self.test_session() as sess:
sess.run(tf.global_variables_initializer())
self.evaluate(regularization)
def _testRegularization(self, type, scale):
tf.reset_default_graph()
x = tf.placeholder_with_default(
np.random.randn(64, 128).astype(np.float32), shape=(None, 128))
x = tf.layers.dense(x, 256)
x = tf.layers.dense(x, 128)
regularization = optim.regularization_penalty(type, scale)
self.assertEqual(0, len(regularization.shape.as_list()))
with self.test_session(tf.get_default_graph()) as sess:
sess.run(tf.global_variables_initializer())
sess.run(regularization)
def testRegulaizationMissingScaleValue(self):
with self.assertRaises(ValueError):
optim.regularization_penalty("l1_l2", 1e-4, weights_list=[])
def testRegulaizationInvalidType(self):
with self.assertRaises(ValueError):
optim.regularization_penalty("l3", 1e-4, weights_list=[])