def test_localNetUpSampleResnetBlock():
"""
Test the layer.LocalNetUpSampleResnetBlock class, its default attributes and its call() function.
"""
batch_size = 5
channels = 4
input_size = (32, 32, 16)
output_size = (64, 64, 32)
nonskip_tensor_size = (batch_size, ) + input_size + (channels, )
skip_tensor_size = (batch_size, ) + output_size + (channels, )
# Test __init__() and build()
model = layer.LocalNetUpSampleResnetBlock(8)
model.build([nonskip_tensor_size, skip_tensor_size])
assert model._filters == 8
assert model._use_additive_upsampling is True
assert isinstance(model._deconv3d_block, type(layer.Deconv3dBlock(8)))
assert isinstance(model._additive_upsampling,
type(layer.AdditiveUpSampling(output_size)))
assert isinstance(model._conv3d_block, type(layer.Conv3dBlock(8)))
assert isinstance(model._residual_block,
type(layer.LocalNetResidual3dBlock(8)))
def test_deconv3dBlock():
"""
Test the layer.Deconv3dBlock class and its default attributes. No need to test the call() function since a
concatenation of tensorflow classes
"""
deconv3dBlock = layer.Deconv3dBlock(8)
assert isinstance(deconv3dBlock._deconv3d, layer.Deconv3d)
assert deconv3dBlock._deconv3d._Conv3DTranspose is None
deconv3dBlock._deconv3d.build((8, 8))
assert isinstance(
deconv3dBlock._deconv3d._Conv3DTranspose,
type(tf.keras.layers.Conv3DTranspose(8, 3)),
)
assert tuple(deconv3dBlock._deconv3d._kernel_size) == (3, 3, 3)
assert tuple(deconv3dBlock._deconv3d._strides) == (1, 1, 1)
assert deconv3dBlock._deconv3d._padding == "same"
assert deconv3dBlock._deconv3d._Conv3DTranspose.use_bias is False
assert isinstance(deconv3dBlock._act._act, type(tf.keras.activations.relu))
assert isinstance(deconv3dBlock._norm._norm,
type(tf.keras.layers.BatchNormalization()))
def __init__(
self,
filters: int,
output_padding: Union[int, Tuple, List],
kernel_size: Union[int, Tuple, List],
padding: str,
strides: Union[int, Tuple, List],
output_shape: Tuple,
name: str = "AdditiveUpsampling",
):
"""
Addictive up-sampling layer.
:param filters: number of channels for output
:param output_padding: padding for output
:param kernel_size: arg for deconv3d
:param padding: arg for deconv3d
:param strides: arg for deconv3d
:param output_shape: shape of the output tensor
:param name: name of the layer.
"""
super().__init__(name=name)
self.deconv3d = layer.Deconv3dBlock(
filters=filters,
output_padding=output_padding,
kernel_size=kernel_size,
strides=strides,
padding=padding,
)
self.resize = layer.Resize3d(shape=output_shape)
def build_up_sampling_block(
self,
filters: int,
output_padding: Union[Tuple[int, ...], int],
kernel_size: Union[Tuple[int, ...], int],
padding: str,
strides: Union[Tuple[int, ...], int],
output_shape: Tuple[int, ...],
) -> Union[tf.keras.Model, tfkl.Layer]:
"""
Build a block for up-sampling.
This block changes the tensor shape (width, height, depth),
but it does not changes the number of channels.
:param filters: number of channels for output
:param output_padding: padding for output
:param kernel_size: arg for deconv3d
:param padding: arg for deconv3d
:param strides: arg for deconv3d
:param output_shape: shape of the output tensor
:return: a block consists of one or multiple layers
"""
return layer.Deconv3dBlock(
filters=filters,
output_padding=output_padding,
kernel_size=kernel_size,
strides=strides,
padding=padding,
)
def test_upsampleResnetBlock():
"""
Test the layer.UpSampleResnetBlock class and its default attributes. No need to test the call() function since a
concatenation of tensorflow classes
"""
batch_size = 5
channels = 4
input_size = (32, 32, 16)
output_size = (64, 64, 32)
input_tensor_size = (batch_size, ) + input_size + (channels, )
skip_tensor_size = (batch_size, ) + output_size + (channels // 2, )
model = layer.UpSampleResnetBlock(8)
model.build([input_tensor_size, skip_tensor_size])
assert model._filters == 8
assert model._concat is False
assert isinstance(model._conv3d_block, type(layer.Conv3dBlock(8)))
assert isinstance(model._residual_block, type(layer.Residual3dBlock(8)))
assert isinstance(model._deconv3d_block, type(layer.Deconv3dBlock(8)))
def test_deconv3d_block():
"""
Test the layer.Deconv3dBlock class and its default attributes.
"""
deconv3d_block = layer.Deconv3dBlock(8)
assert isinstance(deconv3d_block._deconv3d, layer.Deconv3d)
assert deconv3d_block._deconv3d._deconv3d is None
deconv3d_block._deconv3d.build((8, 8))
assert isinstance(
deconv3d_block._deconv3d._deconv3d, tf.keras.layers.Conv3DTranspose
)
assert tuple(deconv3d_block._deconv3d._kernel_size) == (3, 3, 3)
assert tuple(deconv3d_block._deconv3d._strides) == (1, 1, 1)
assert deconv3d_block._deconv3d._padding == "same"
assert deconv3d_block._deconv3d._deconv3d.use_bias is False
assert isinstance(deconv3d_block._act._act, type(tf.keras.activations.relu))
assert isinstance(deconv3d_block._norm._norm, tf.keras.layers.BatchNormalization)