def testReuseVariables(self):
height, width = 3, 3
images = random_ops.random_uniform((5, height, width, 4), seed=1)
normalization_ops.layer_norm(images, scale=True, scope='IN')
normalization_ops.layer_norm(images, scale=True, scope='IN', reuse=True)
beta = get_variables('beta')
gamma = get_variables('gamma')
self.assertEqual(1, len(beta))
self.assertEqual(1, len(gamma))
def testValueCorrectWithReuseVars(self):
height, width = 3, 3
image_shape = (10, height, width, 4)
images = random_ops.random_uniform(image_shape, seed=1)
output_train = normalization_ops.layer_norm(images, scope='IN')
output_eval = normalization_ops.layer_norm(images, scope='IN', reuse=True)
with self.cached_session() as sess:
sess.run(variables.global_variables_initializer())
# output_train and output_eval should be the same.
train_np, eval_np = sess.run([output_train, output_eval])
self.assertAllClose(train_np, eval_np)
def testCreateVariables_NCHW(self):
height, width = 3, 3
images = random_ops.random_uniform((5, 2, height, width), seed=1)
normalization_ops.layer_norm(images,
channels_axis=-3,
center=True,
scale=True)
beta = get_variables('beta')[0]
gamma = get_variables('gamma')[0]
self.assertEqual('LayerNorm/beta', beta.op.name)
self.assertEqual('LayerNorm/gamma', gamma.op.name)
def doOutputTest(self,
input_shape,
channels_axis=None,
reduction_axes=None,
tol=1e-1):
# Select the axis for the channel and the dimensions along which statistics
# are accumulated.
if channels_axis < 0:
channels_axis += len(input_shape)
reduced_axes = [channels_axis + 1]
for a in reduction_axes:
if a < 0:
a += len(input_shape)
if a < channels_axis:
reduced_axes.append(a)
else:
reduced_axes.append(a + 1)
reduced_axes = tuple(reduced_axes)
channels = input_shape[channels_axis]
# Calculate the final shape for the output Tensor.
axes_before_channels = input_shape[:channels_axis]
axes_after_channels = input_shape[channels_axis + 1:]
outputs_shape = (axes_before_channels + [1, channels] +
axes_after_channels)
# Calculate the final shape for the output statistics.
reduced_shape = []
for i, a in enumerate(outputs_shape):
if i not in reduced_axes:
reduced_shape.append(a)
mu = 1.0
sigma = 1.0
# Determine shape of Tensor after normalization.
expected_mean = np.zeros(reduced_shape)
expected_var = np.ones(reduced_shape)
inputs = random_ops.random_normal(input_shape, seed=0) * sigma + mu
output_op = normalization_ops.layer_norm(inputs,
center=False,
scale=False,
channels_axis=channels_axis,
training=True)
with self.cached_session() as sess:
sess.run(variables.global_variables_initializer())
outputs = sess.run(output_op)
# Make sure that there are no NaNs
self.assertFalse(np.isnan(outputs).any())
outputs = np.swapaxes(outputs, 0, channels_axis)
outputs = np.reshape(outputs, outputs_shape)
mean = np.mean(outputs, axis=reduced_axes, dtype=np.float32)
var = np.var(outputs, axis=reduced_axes, dtype=np.float32)
# The mean and variance of each example should be close to 0 and 1
# respectively.
self.assertAllClose(expected_mean, mean, rtol=tol, atol=tol)
self.assertAllClose(expected_var, var, rtol=tol, atol=tol)
def testCreateOpNoScaleCenter(self):
height, width = 3, 3
images = random_ops.random_uniform((5, height, width, 3),
dtype=dtypes.float32,
seed=1)
output = normalization_ops.layer_norm(images, center=False, scale=False)
self.assertListEqual([5, height, width, 3], output.shape.as_list())
self.assertEqual(0, len(get_variables('beta')))
self.assertEqual(0, len(get_variables('gamma')))
def testCreateOp(self): height, width = 3, 3 images = random_ops.random_uniform((5, height, width, 2), seed=1) output = normalization_ops.layer_norm(images, channels_axis=-1) self.assertListEqual([5, height, width, 2], output.shape.as_list())
def testParamsShapeNotFullyDefinedChannelsAxis(self):
inputs = array_ops.placeholder(dtypes.float32, shape=(1, 3, 4, None))
with self.assertRaisesRegex(ValueError, 'undefined channel dimension'):
normalization_ops.layer_norm(inputs, channels_axis=-1)
def testUnknownShape(self):
inputs = array_ops.placeholder(dtypes.float32)
with self.assertRaisesRegex(ValueError, 'undefined rank'):
normalization_ops.layer_norm(inputs)
def testAxisIsBad(self):
inputs = array_ops.placeholder(dtypes.float32, shape=(1, 2, 4, 5))
with self.assertRaisesRegex(ValueError, 'Axis is out of bounds.'):
normalization_ops.layer_norm(inputs, channels_axis=5)
def func(a, b):
x = tf.matmul(a, b)
x = normalization_ops.layer_norm(x)
x = nn_ops.gelu(x)
return x
