def testFunctionalConv2DTransposeNoReuse(self): height, width = 7, 9 images = random_ops.random_uniform((5, height, width, 3), seed=1) conv_layers.conv2d_transpose(images, 32, [3, 3]) self.assertEqual(len(variables.trainable_variables()), 2) conv_layers.conv2d_transpose(images, 32, [3, 3]) self.assertEqual(len(variables.trainable_variables()), 4)
def testInvalidKernelSize(self):
height, width = 7, 9
images = random_ops.random_uniform((5, height, width, 3), seed=1)
with self.assertRaisesRegexp(ValueError, 'kernel_size'):
conv_layers.conv2d_transpose(images, 32, (1, 2, 3))
with self.assertRaisesRegexp(ValueError, 'kernel_size'):
conv_layers.conv2d_transpose(images, 32, None)
def testInvalidStrides(self):
height, width = 7, 9
images = random_ops.random_uniform((5, height, width, 3), seed=1)
with self.assertRaisesRegexp(ValueError, 'strides'):
conv_layers.conv2d_transpose(images, 32, 3, strides=(1, 2, 3))
with self.assertRaisesRegexp(ValueError, 'strides'):
conv_layers.conv2d_transpose(images, 32, 3, strides=None)
def testFunctionalConv2DTransposeReuseFromScope(self):
with variable_scope.variable_scope('scope'):
height, width = 7, 9
images = random_ops.random_uniform((5, height, width, 3), seed=1)
conv_layers.conv2d_transpose(images, 32, [3, 3], name='deconv1')
self.assertEqual(len(variables.trainable_variables()), 2)
with variable_scope.variable_scope('scope', reuse=True):
conv_layers.conv2d_transpose(images, 32, [3, 3], name='deconv1')
self.assertEqual(len(variables.trainable_variables()), 2)
def testFunctionalConv2DTransposeNoReuse(self):
height, width = 7, 9
images = random_ops.random_uniform((5, height, width, 3), seed=1)
conv_layers.conv2d_transpose(images, 32, [3, 3])
self.assertEqual(
len(ops.get_collection(ops.GraphKeys.TRAINABLE_VARIABLES)), 2)
conv_layers.conv2d_transpose(images, 32, [3, 3])
self.assertEqual(
len(ops.get_collection(ops.GraphKeys.TRAINABLE_VARIABLES)), 4)
def testConv2DTransposeFloat16(self):
height, width = 7, 9
images = random_ops.random_uniform((5, height, width, 4), dtype='float16')
output = conv_layers.conv2d_transpose(images, 32, [3, 3],
activation=nn_ops.relu)
self.assertListEqual(output.get_shape().as_list(),
[5, height + 2, width + 2, 32])
def testFunctionalConv2DTransposeInitializerFromScope(self):
with self.test_session() as sess:
with variable_scope.variable_scope(
'scope', initializer=init_ops.ones_initializer()):
height, width = 7, 9
images = random_ops.random_uniform((5, height, width, 3), seed=1)
conv_layers.conv2d_transpose(images, 32, [3, 3], name='deconv1')
weights = variables.trainable_variables()
# Check the names of weights in order.
self.assertTrue('kernel' in weights[0].name)
self.assertTrue('bias' in weights[1].name)
sess.run(variables.global_variables_initializer())
weights = sess.run(weights)
# Check that the kernel weights got initialized to ones (from scope)
self.assertAllClose(weights[0], np.ones((3, 3, 32, 3)))
# Check that the bias still got initialized to zeros.
self.assertAllClose(weights[1], np.zeros((32)))
def testInvalidDataFormat(self):
height, width = 7, 9
images = random_ops.random_uniform((5, height, width, 3), seed=1)
with self.assertRaisesRegexp(ValueError, 'data_format'):
conv_layers.conv2d_transpose(images, 32, 3, data_format='invalid')
