def test_multi_class(self):
self.dnn = KerasDnnClassifier(multi_class=True, loss="")
self.assertEqual(self.dnn.loss, "categorical_crossentropy")
self.assertEqual(self.dnn.multi_class, True)
self.dnn.multi_class = False
self.assertEqual(self.dnn.loss, "binary_crossentropy")
self.assertEqual(self.dnn.multi_class, False)
with self.assertRaises(ValueError):
self.dnn = KerasDnnClassifier(multi_class=True, loss='hinge')
with self.assertRaises(ValueError):
self.dnn = KerasDnnClassifier(multi_class=False,
loss='kullback_leibler_divergence')
def test_parameter_length(self):
self.dnn = KerasDnnClassifier(hidden_layer_sizes=[1, 2, 3, 4, 5, 6],
dropout_rate=0.2,
activations='tanh')
self.assertEqual(self.dnn.dropout_rate, [0.2] * 6)
self.assertEqual(self.dnn.activations, ['tanh'] * 6)
self.dnn.dropout_rate = [0.1, 0.2, 0.3, 0.4, 0.5, 0.6]
self.assertEqual(self.dnn.dropout_rate, [0.1, 0.2, 0.3, 0.4, 0.5, 0.6])
self.dnn.activations = ["tanh", "sigmoid"] * 3
self.assertEqual(self.dnn.activations, ["tanh", "sigmoid"] * 3)
self.dnn.dropout_rate = 0.4
self.assertEqual(self.dnn.dropout_rate, [0.4] * 6)
with self.assertRaises(ValueError):
self.dnn.activations = "roundabout"
with self.assertRaises(ValueError):
self.dnn.hidden_layer_sizes = [1, 2, 3]
with self.assertRaises(ValueError):
self.dnn.dropout_rate = [0.2, 0.6]
def setUp(self):
self.model_wrapper = KerasDnnClassifier()
self.dnn = self.model_wrapper