def test_lstm_hourglass_checks_compression_factor(self):
"""
Test that lstm_hourglass validates parameter requirements
"""
with self.assertRaises(ValueError):
lstm_hourglass(3, compression_factor=2)
with self.assertRaises(ValueError):
lstm_hourglass(3, compression_factor=-1)
def test_lstm_hourglass_basic(self):
"""
Test that lstm_hourglass implements the correct parameters
"""
model = lstm_hourglass(
3,
func="tanh",
out_func="relu",
optimizer="sgd",
optimizer_kwargs={
"lr": 0.02,
"momentum": 0.001
},
compile_kwargs={"loss": "mae"},
)
# Ensure that the input dimension to Keras model matches the number of features.
self.assertEqual(model.layers[0].input_shape[2], 3)
# Ensure that the dimension of each encoding layer matches the expected dimension.
self.assertEqual([model.layers[i].input_shape[2] for i in range(1, 4)],
[3, 2, 2])
# Ensure that the dimension of each decoding layer (excluding last decoding layer)
# matches the expected dimension.
self.assertEqual([model.layers[i].input_shape[2] for i in range(4, 6)],
[2, 2])
# Ensure that the dimension of last decoding layer matches the expected dimension.
self.assertEqual(model.layers[6].input_shape[1], 3)
# Ensure activation functions in the encoding part (layers 0-2)
# match expected activation functions
self.assertEqual(
[model.layers[i].activation.__name__ for i in range(0, 3)],
["tanh", "tanh", "tanh"],
)
# Ensure activation functions in the decoding part (layers 3-5)
# match expected activation functions
self.assertEqual(
[model.layers[i].activation.__name__ for i in range(3, 6)],
["tanh", "tanh", "tanh"],
)
# Ensure activation function for the output layer matches expected activation function
self.assertEqual(model.layers[6].activation.__name__, "relu")
# Assert that the expected Keras optimizer is used
self.assertEqual(model.optimizer.__class__, optimizers.SGD)
# Assert equality of difference up to 7 decimal places
# Note that AlmostEquality is used as Keras can use a value approximately equal
# to the given parameter rather than the exact value.
self.assertAlmostEqual(K.eval(model.optimizer.lr), 0.02)
self.assertAlmostEqual(K.eval(model.optimizer.momentum), 0.001)
# Assert that the correct loss function is used.
self.assertEqual(model.loss, "mae")
def test_lstm_hourglass_basic(self):
"""
Test that lstm_hourglass implements the correct parameters
"""
model = lstm_hourglass(
n_features=3,
func="tanh",
out_func="relu",
optimizer="SGD",
optimizer_kwargs={
"lr": 0.02,
"momentum": 0.001
},
compile_kwargs={"loss": "mae"},
)
# Ensure that the input dimension to Keras model matches the number of features.
self.assertEqual(model.layers[0].input_shape[2], 3)
# Ensure that the dimension of each encoding layer matches the expected dimension.
self.assertEqual([model.layers[i].input_shape[2] for i in range(1, 4)],
[3, 2, 2])
# Ensure that the dimension of each decoding layer (excluding last decoding layer)
# matches the expected dimension.
self.assertEqual([model.layers[i].input_shape[2] for i in range(4, 6)],
[2, 2])
# Ensure that the dimension of last decoding layer matches the expected dimension.
self.assertEqual(model.layers[6].input_shape[1], 3)
# Ensure activation functions in the encoding part (layers 0-2)
# match expected activation functions
self.assertEqual(
[model.layers[i].activation.__name__ for i in range(0, 3)],
["tanh", "tanh", "tanh"],
)
# Ensure activation functions in the decoding part (layers 3-5)
# match expected activation functions
self.assertEqual(
[model.layers[i].activation.__name__ for i in range(3, 6)],
["tanh", "tanh", "tanh"],
)
# Ensure activation function for the output layer matches expected activation function
self.assertEqual(model.layers[6].activation.__name__, "relu")
# Assert that the expected Keras optimizer is used
self.assertEqual(model.optimizer.__class__, optimizers.SGD)
# Assert that the correct loss function is used.
self.assertEqual(model.loss, "mae")
def test_lstm_defaults(self):
"""
Test that models with all defaults are created without failure and are equal
"""
# Create models with default parameters
base = lstm_model(4)
symmetric = lstm_symmetric(4)
hourglass = lstm_hourglass(4)
# Ensure LSTM model layers are equal to base model layers
for i in range(len(base.layers)):
# Rename layers so as to not fail on names, only configuration
config = base.layers[i].get_config().update({"name": "test"})
symmetric_config = symmetric.layers[i].get_config().update(
{"name": "test"})
hourglass_config = hourglass.layers[i].get_config().update(
{"name": "test"})
assert config == symmetric_config
assert config == hourglass_config
def test_lstm_hourglass_checks_enc_layers(self):
"""
Test that lstm_hourglass validates parameter requirements
"""
with self.assertRaises(ValueError):
lstm_hourglass(3, encoding_layers=0)
