def test_vector_input_raises_error(self):
x = tf.random.uniform((3, ), dtype=tf.float32)
with self.assertRaises(ValueError):
gdn.GDN(inverse=False, rectify=False,
data_format="channels_last")(x)
with self.assertRaises(ValueError):
gdn.GDN(inverse=True, rectify=True,
data_format="channels_first")(x)
def test_variables_are_not_enumerated_when_overridden(self):
layer = gdn.GDN()
layer.beta_parameter = [1]
layer.gamma_parameter = [.1]
layer.build((None, 1))
self.assertEmpty(layer.weights)
self.assertEmpty(layer.trainable_weights)
def test_channels_first_has_correct_output(self, rank): # This tests that the layer produces the correct output for a number of # different input dimensionalities with 'channels_first' data format. x = tf.random.uniform((6, 5, 4, 3, 2, 1)[:rank], dtype=tf.float32) y = gdn.GDN(inverse=False, rectify=False, data_format="channels_first")(x) self.assertEqual(x.shape, y.shape) self.assertAllClose(y, x / (1 + .1 * abs(x)), rtol=0, atol=1e-6)
def test_quadratic_gdn_has_correct_output(self):
x = tf.random.uniform((1, 2, 3, 4), -.5, .5, dtype=tf.float32)
y = gdn.GDN(
inverse=False, rectify=False,
alpha_parameter=2, epsilon_parameter=.5)(x)
self.assertEqual(x.shape, y.shape)
self.assertAllClose(y, x / tf.sqrt(1 + .1 * (x ** 2)), rtol=0, atol=1e-6)
def test_attributes_cannot_be_set_after_build(self):
layer = gdn.GDN()
layer.build((None, 2))
with self.assertRaises(RuntimeError):
layer.inverse = True
with self.assertRaises(RuntimeError):
layer.rectify = True
with self.assertRaises(RuntimeError):
layer.data_format = "channels_first"
with self.assertRaises(RuntimeError):
layer.alpha_parameter = 5
with self.assertRaises(RuntimeError):
layer.beta_parameter = tf.ones((5, ))
with self.assertRaises(RuntimeError):
layer.gamma_parameter = tf.ones((5, 5))
with self.assertRaises(RuntimeError):
layer.epsilon_parameter = 1 / 3
with self.assertRaises(RuntimeError):
layer.alpha_initializer = "ones"
with self.assertRaises(RuntimeError):
layer.beta_initializer = tf.keras.initializers.Ones()
with self.assertRaises(RuntimeError):
layer.gamma_initializer = tf.keras.initializers.Ones()
with self.assertRaises(RuntimeError):
layer.epsilon_initializer = "ones"
def test_variables_receive_gradients(self):
x = tf.random.uniform((1, 2), dtype=tf.float32)
layer = gdn.GDN(inverse=False, rectify=True)
with tf.GradientTape() as g:
y = layer(x)
grads = g.gradient(y, layer.trainable_variables)
self.assertLen(grads, 2)
self.assertNotIn(None, grads)
def test_fixed_gdn_has_correct_output(self):
x = tf.random.uniform((10, 3), -.5, .5, dtype=tf.float32)
y = gdn.GDN(
inverse=False, rectify=False,
beta_parameter=[0, 0, 0], gamma_parameter=tf.ones((3, 3)))(x)
self.assertEqual(x.shape, y.shape)
expected_y = x / tf.reduce_sum(abs(x), axis=-1, keepdims=True)
self.assertAllClose(y, expected_y, rtol=0, atol=1e-6)
def _run_gdn(self, x, shape, inverse, rectify, data_format):
inputs = tf.placeholder(tf.float32, shape)
layer = gdn.GDN(
inverse=inverse, rectify=rectify, data_format=data_format)
outputs = layer(inputs)
with self.test_session() as sess:
tf.global_variables_initializer().run()
y, = sess.run([outputs], {inputs: x})
return y
def test_dtypes_are_correct_with_mixed_precision(self):
tf.keras.mixed_precision.set_global_policy("mixed_float16")
try:
x = tf.random.uniform((4, 3), dtype=tf.float16)
layer = gdn.GDN()
y = layer(x)
for variable in layer.variables:
self.assertEqual(variable.dtype, tf.float32)
self.assertEqual(y.dtype, tf.float16)
finally:
tf.keras.mixed_precision.set_global_policy(None)
def test_variables_receive_gradients(self):
x = tf.random.uniform((1, 2), dtype=tf.float32)
layer = gdn.GDN(inverse=False, rectify=True)
with tf.GradientTape() as g:
y = layer(x)
grads = g.gradient(y, layer.trainable_weights)
self.assertLen(grads, 2)
self.assertNotIn(None, grads)
grad_shapes = [tuple(g.shape) for g in grads]
weight_shapes = [tuple(w.shape) for w in layer.trainable_weights]
self.assertSameElements(grad_shapes, weight_shapes)
def test_variables_are_enumerated(self):
layer = gdn.GDN()
layer.alpha_parameter = None
layer.epsilon_parameter = None
layer.build((None, 5))
self.assertLen(layer.weights, 4)
self.assertLen(layer.trainable_weights, 4)
weight_names = [w.name for w in layer.weights]
self.assertSameElements(weight_names, [
"reparam_alpha:0", "reparam_beta:0", "reparam_gamma:0",
"reparam_epsilon:0"])
def test_channels_last_has_correct_output(self):
# This tests that the layer produces the correct output for a number of
# different input dimensionalities with 'channels_last' data format.
for ndim in [2, 3, 4, 5, 6]:
x = tf.random.uniform((1, 2, 3, 4, 5, 6)[:ndim], dtype=tf.float32)
y = gdn.GDN(inverse=False,
rectify=False,
data_format="channels_last")(x)
self.assertEqual(x.shape, y.shape)
self.assertAllClose(y,
x / tf.sqrt(1 + .1 * (x**2)),
rtol=0,
atol=1e-6)
def test_can_be_saved_within_functional_model(self, build):
inputs = tf.keras.Input(shape=(5, ))
outputs = gdn.GDN()(inputs)
model = tf.keras.Model(inputs=inputs, outputs=outputs)
layer = model.get_layer("gdn")
with self.subTest(name="layer_created_as_expected"):
self.assertIsInstance(layer, gdn.GDN)
self.assertIsInstance(layer.alpha_parameter, tf.Tensor)
self.assertEmpty(layer.alpha_parameter.shape)
self.assertIsInstance(layer.beta_parameter,
parameters.GDNParameter)
self.assertIsInstance(layer.gamma_parameter,
parameters.GDNParameter)
self.assertIsInstance(layer.epsilon_parameter, tf.Tensor)
self.assertEmpty(layer.epsilon_parameter.shape)
if build:
x = tf.random.uniform((5, 5), dtype=tf.float32)
y = model(x)
weight_names = [w.name for w in model.weights]
tempdir = self.create_tempdir()
model_path = os.path.join(tempdir, "model")
# This should force the model to be reconstructed via configs.
model.save(model_path, save_traces=False)
model = tf.keras.models.load_model(model_path)
layer = model.get_layer("gdn")
with self.subTest(name="layer_recreated_as_expected"):
self.assertIsInstance(layer, gdn.GDN)
self.assertIsInstance(layer.alpha_parameter, tf.Tensor)
self.assertEmpty(layer.alpha_parameter.shape)
self.assertIsInstance(layer.beta_parameter,
parameters.GDNParameter)
self.assertIsInstance(layer.gamma_parameter,
parameters.GDNParameter)
self.assertIsInstance(layer.epsilon_parameter, tf.Tensor)
self.assertEmpty(layer.epsilon_parameter.shape)
if build:
with self.subTest(name="model_outputs_identical"):
self.assertAllEqual(model(x), y)
with self.subTest(name="model_weights_identical"):
self.assertSameElements(weight_names,
[w.name for w in model.weights])
def test_rgdn_has_correct_output(self):
x = tf.random.uniform((1, 2, 3, 4), -.5, .5, dtype=tf.float32)
y = gdn.GDN(inverse=False, rectify=True)(x)
self.assertEqual(x.shape, y.shape)
x = tf.maximum(x, 0)
self.assertAllClose(y, x / (1 + .1 * x), rtol=0, atol=1e-6)
def test_igdn_has_correct_output(self):
x = tf.random.uniform((1, 2, 3, 4), dtype=tf.float32)
y = gdn.GDN(inverse=True, rectify=False)(x)
self.assertEqual(x.shape, y.shape)
self.assertAllClose(y, x * (1 + .1 * abs(x)), rtol=0, atol=1e-6)
def test_vector_input_raises_error(self, data_format):
x = tf.zeros((3, ), dtype=tf.float32)
with self.assertRaises(ValueError):
gdn.GDN(data_format=data_format)(x)
def test_invalid_data_format_raises_error(self):
with self.assertRaises(ValueError):
gdn.GDN(data_format="NHWC")
def test_variables_trainable_state_follows_layer(self):
layer = gdn.GDN()
layer.trainable = False
layer.build((None, 1))
self.assertLen(layer.weights, 2)
self.assertEmpty(layer.trainable_weights)
def test_invalid_data_format_raises_error(self):
x = tf.random.uniform((1, 2, 3, 4), dtype=tf.float32)
with self.assertRaises(ValueError):
gdn.GDN(inverse=False, rectify=False, data_format="NHWC")(x)
