class TestPerImageStandardize(CleverHansTest):
def setUp(self):
super(TestPerImageStandardize, self).setUp()
self.input_shape = (128, 32, 32, 3)
self.sess = tf.Session()
self.model = MLP(input_shape=self.input_shape,
layers=[PerImageStandardize(name='output')])
self.x = tf.placeholder(shape=self.input_shape, dtype=tf.float32)
self.y = self.model.get_layer(self.x, 'output')
self.y_true = tf.map_fn(
lambda ex: tf.image.per_image_standardization(ex), self.x)
def run_and_check_output(self, x):
y, y_true = self.sess.run([self.y, self.y_true], feed_dict={self.x: x})
self.assertClose(y, y_true)
def test_random_inputs(self):
x = np.random.rand(*self.input_shape)
self.run_and_check_output(x)
def test_uniform_inputs(self):
x = np.ones(self.input_shape)
self.run_and_check_output(x)
def test_no_drop():
"""test_no_drop: Make sure dropout does nothing by default (so it does not
cause stochasticity at test time)"""
model = MLP(input_shape=[1, 1], layers=[Dropout(name='output')])
x = tf.constant([[1]], dtype=tf.float32)
y = model.get_layer(x, 'output')
sess = tf.Session()
# Do multiple runs because dropout is stochastic
for _ in range(10):
y_value = sess.run(y)
assert y_value == 1.
def test_drop():
"""test_drop: Make sure dropout is activated successfully"""
# We would like to configure the test to deterministically drop,
# so that the test does not need to use multiple runs.
# However, tf.nn.dropout divides by include_prob, so zero or
# infinitesimal include_prob causes NaNs.
# 1e-8 does not cause NaNs and shouldn't be a significant source
# of test flakiness relative to dependency downloads failing, etc.
model = MLP(input_shape=[1, 1],
layers=[Dropout(name='output', include_prob=1e-8)])
x = tf.constant([[1]], dtype=tf.float32)
y = model.get_layer(x, 'output', dropout=True)
sess = tf.Session()
y_value = sess.run(y)
# Subject to very rare random failure because include_prob is not exact 0
assert y_value == 0., y_value
def test_override():
# Make sure dropout_dict changes dropout probabilities successful
# We would like to configure the test to deterministically drop,
# so that the test does not need to use multiple runs.
# However, tf.nn.dropout divides by include_prob, so zero or
# infinitesimal include_prob causes NaNs.
# For this test, random failure to drop will not cause the test to fail.
# The stochastic version should not even run if everything is working
# right.
model = MLP(input_shape=[1, 1], layers=[Dropout(name='output',
include_prob=1e-8)])
x = tf.constant([[1]], dtype=tf.float32)
dropout_dict = {'output': 1.}
y = model.get_layer(x, 'output', dropout=True, dropout_dict=dropout_dict)
sess = tf.Session()
y_value = sess.run(y)
assert y_value == 1., y_value
class TestPerImageStandardize(CleverHansTest):
"""
Tests for the PerImageStandardize class.
"""
def setUp(self):
"""
Set up session and build model graph
"""
super(TestPerImageStandardize, self).setUp()
self.input_shape = (128, 32, 32, 3)
self.sess = tf.compat.v1.Session()
self.model = MLP(input_shape=self.input_shape,
layers=[PerImageStandardize(name='output')])
self.x = tf.compat.v1.placeholder(shape=self.input_shape,
dtype=tf.float32)
self.y = self.model.get_layer(self.x, 'output')
self.y_true = tf.map_fn(tf.image.per_image_standardization, self.x)
def run_and_check_output(self, x):
"""
Make sure y and y_true evaluate to the same value
"""
y, y_true = self.sess.run([self.y, self.y_true], feed_dict={self.x: x})
self.assertClose(y, y_true)
def test_random_inputs(self):
"""
Test on random inputs
"""
x = np.random.rand(*self.input_shape)
self.run_and_check_output(x)
def test_ones_inputs(self):
"""
Test with input set to all ones.
"""
x = np.ones(self.input_shape)
self.run_and_check_output(x)