def create_model_brick():
layers = [
conv_brick(2, 1, 64),
bn_brick(),
LeakyRectifier(leak=LEAK),
conv_brick(7, 2, 128),
bn_brick(),
LeakyRectifier(leak=LEAK),
conv_brick(5, 2, 256),
bn_brick(),
LeakyRectifier(leak=LEAK),
conv_brick(7, 2, 256),
bn_brick(),
LeakyRectifier(leak=LEAK),
conv_brick(4, 1, 512),
bn_brick(),
LeakyRectifier(leak=LEAK),
conv_brick(1, 1, 2 * NLAT)
]
encoder_mapping = ConvolutionalSequence(layers=layers,
num_channels=NUM_CHANNELS,
image_size=IMAGE_SIZE,
use_bias=False,
name='encoder_mapping')
encoder = GaussianConditional(encoder_mapping, name='encoder')
layers = [
conv_transpose_brick(4, 1, 512),
bn_brick(),
LeakyRectifier(leak=LEAK),
conv_transpose_brick(7, 2, 256),
bn_brick(),
LeakyRectifier(leak=LEAK),
conv_transpose_brick(5, 2, 256),
bn_brick(),
LeakyRectifier(leak=LEAK),
conv_transpose_brick(7, 2, 128),
bn_brick(),
LeakyRectifier(leak=LEAK),
conv_transpose_brick(2, 1, 64),
bn_brick(),
LeakyRectifier(leak=LEAK),
conv_brick(1, 1, NUM_CHANNELS),
Logistic()
]
decoder_mapping = ConvolutionalSequence(layers=layers,
num_channels=NLAT,
image_size=(1, 1),
use_bias=False,
name='decoder_mapping')
decoder = DeterministicConditional(decoder_mapping, name='decoder')
layers = [
conv_brick(2, 1, 64),
LeakyRectifier(leak=LEAK),
conv_brick(7, 2, 128),
bn_brick(),
LeakyRectifier(leak=LEAK),
conv_brick(5, 2, 256),
bn_brick(),
LeakyRectifier(leak=LEAK),
conv_brick(7, 2, 256),
bn_brick(),
LeakyRectifier(leak=LEAK),
conv_brick(4, 1, 512),
bn_brick(),
LeakyRectifier(leak=LEAK)
]
x_discriminator = ConvolutionalSequence(layers=layers,
num_channels=NUM_CHANNELS,
image_size=IMAGE_SIZE,
use_bias=False,
name='x_discriminator')
x_discriminator.push_allocation_config()
layers = [
conv_brick(1, 1, 1024),
LeakyRectifier(leak=LEAK),
conv_brick(1, 1, 1024),
LeakyRectifier(leak=LEAK)
]
z_discriminator = ConvolutionalSequence(layers=layers,
num_channels=NLAT,
image_size=(1, 1),
use_bias=False,
name='z_discriminator')
z_discriminator.push_allocation_config()
layers = [
conv_brick(1, 1, 2048),
LeakyRectifier(leak=LEAK),
conv_brick(1, 1, 2048),
LeakyRectifier(leak=LEAK),
conv_brick(1, 1, 1)
]
joint_discriminator = ConvolutionalSequence(
layers=layers,
num_channels=(x_discriminator.get_dim('output')[0] +
z_discriminator.get_dim('output')[0]),
image_size=(1, 1),
name='joint_discriminator')
discriminator = XZJointDiscriminator(x_discriminator,
z_discriminator,
joint_discriminator,
name='discriminator')
ali = ALI(encoder,
decoder,
discriminator,
weights_init=GAUSSIAN_INIT,
biases_init=ZERO_INIT,
name='ali')
ali.push_allocation_config()
encoder_mapping.layers[-1].use_bias = True
encoder_mapping.layers[-1].tied_biases = False
decoder_mapping.layers[-2].use_bias = True
decoder_mapping.layers[-2].tied_biases = False
x_discriminator.layers[0].use_bias = True
x_discriminator.layers[0].tied_biases = True
ali.initialize()
raw_marginals, = next(
create_celeba_data_streams(500, 500)[0].get_epoch_iterator())
b_value = get_log_odds(raw_marginals)
decoder_mapping.layers[-2].b.set_value(b_value)
return ali
def create_model_brick():
encoder_mapping = MLP(
dims=[2 * INPUT_DIM, GEN_HIDDEN, GEN_HIDDEN, NLAT],
activations=[
Sequence([
BatchNormalization(GEN_HIDDEN).apply,
GEN_ACTIVATION().apply
],
name='encoder_h1'),
Sequence([
BatchNormalization(GEN_HIDDEN).apply,
GEN_ACTIVATION().apply
],
name='encoder_h2'),
Identity(name='encoder_out')
],
use_bias=False,
name='encoder_mapping')
encoder = COVConditional(encoder_mapping, (INPUT_DIM, ),
name='encoder')
decoder_mapping = MLP(dims=[
NLAT, GEN_HIDDEN, GEN_HIDDEN, GEN_HIDDEN, GEN_HIDDEN, INPUT_DIM
],
activations=[
Sequence([
BatchNormalization(GEN_HIDDEN).apply,
GEN_ACTIVATION().apply
],
name='decoder_h1'),
Sequence([
BatchNormalization(GEN_HIDDEN).apply,
GEN_ACTIVATION().apply
],
name='decoder_h2'),
Sequence([
BatchNormalization(GEN_HIDDEN).apply,
GEN_ACTIVATION().apply
],
name='decoder_h3'),
Sequence([
BatchNormalization(GEN_HIDDEN).apply,
GEN_ACTIVATION().apply
],
name='decoder_h4'),
Identity(name='decoder_out')
],
use_bias=False,
name='decoder_mapping')
decoder = DeterministicConditional(decoder_mapping, name='decoder')
x_discriminator = Identity(name='x_discriminator')
z_discriminator = Identity(name='z_discriminator')
joint_discriminator = Sequence(application_methods=[
LinearMaxout(input_dim=INPUT_DIM + NLAT,
output_dim=DISC_HIDDEN,
num_pieces=MAXOUT_PIECES,
weights_init=GAUSSIAN_INIT,
biases_init=ZERO_INIT,
name='discriminator_h1').apply,
LinearMaxout(input_dim=DISC_HIDDEN,
output_dim=DISC_HIDDEN,
num_pieces=MAXOUT_PIECES,
weights_init=GAUSSIAN_INIT,
biases_init=ZERO_INIT,
name='discriminator_h2').apply,
LinearMaxout(input_dim=DISC_HIDDEN,
output_dim=DISC_HIDDEN,
num_pieces=MAXOUT_PIECES,
weights_init=GAUSSIAN_INIT,
biases_init=ZERO_INIT,
name='discriminator_h3').apply,
Linear(input_dim=DISC_HIDDEN,
output_dim=1,
weights_init=GAUSSIAN_INIT,
biases_init=ZERO_INIT,
name='discriminator_out').apply
],
name='joint_discriminator')
discriminator = XZJointDiscriminator(x_discriminator,
z_discriminator,
joint_discriminator,
name='discriminator')
ali = ALI(encoder=encoder,
decoder=decoder,
discriminator=discriminator,
weights_init=GAUSSIAN_INIT,
biases_init=ZERO_INIT,
name='ali')
ali.push_allocation_config()
encoder_mapping.linear_transformations[-1].use_bias = True
decoder_mapping.linear_transformations[-1].use_bias = True
ali.initialize()
print("Number of parameters in discriminator: {}".format(
numpy.sum([
numpy.prod(v.shape.eval()) for v in Selector(
ali.discriminator).get_parameters().values()
])))
print("Number of parameters in encoder: {}".format(
numpy.sum([
numpy.prod(v.shape.eval())
for v in Selector(ali.encoder).get_parameters().values()
])))
print("Number of parameters in decoder: {}".format(
numpy.sum([
numpy.prod(v.shape.eval())
for v in Selector(ali.decoder).get_parameters().values()
])))
return ali