def _build_block(self, hp, output_node, block_type):
if block_type == RESNET:
return basic.ResNetBlock().build(hp, output_node)
elif block_type == XCEPTION:
return basic.XceptionBlock().build(hp, output_node)
elif block_type == VANILLA:
return basic.ConvBlock().build(hp, output_node)
def build(self, hp, inputs=None):
input_node = nest.flatten(inputs)[0]
output_node = input_node
block_type = self.block_type or hp.Choice(
'block_type', ['resnet', 'xception', 'vanilla'], default='vanilla')
normalize = self.normalize
if normalize is None:
normalize = hp.Boolean('normalize', default=False)
augment = self.augment
if augment is None:
augment = hp.Boolean('augment', default=False)
if normalize:
output_node = preprocessing.Normalization().build(hp, output_node)
if augment:
output_node = preprocessing.ImageAugmentation().build(
hp, output_node)
if block_type == 'resnet':
output_node = basic.ResNetBlock().build(hp, output_node)
elif block_type == 'xception':
output_node = basic.XceptionBlock().build(hp, output_node)
elif block_type == 'vanilla':
output_node = basic.ConvBlock().build(hp, output_node)
return output_node
def test_resnet_block(init, build):
utils.block_basic_exam(
basic.ResNetBlock(),
tf.keras.Input(shape=(32, 32, 3), dtype=tf.float32),
['version', 'pooling'],
)
assert init.called
assert build.called
def test_resnet_invalid_kwargs2():
with pytest.raises(ValueError) as info:
basic.ResNetBlock(input_shape=(10,))
assert 'Argument "input_shape" is not' in str(info.value)
def test_resnet_invalid_kwargs():
with pytest.raises(ValueError) as info:
basic.ResNetBlock(include_top=True)
assert 'Argument "include_top" is not' in str(info.value)
def test_resnet_block():
utils.block_basic_exam(
basic.ResNetBlock(),
tf.keras.Input(shape=(32, 32, 3), dtype=tf.float32),
['version', 'pooling'],
)