def generator_outputs(inputs, sizes):
latent = inputs[0]
input_energy = inputs[1]
h = Lambda(lambda x: x[0] * x[1])([latent, scale(input_energy, 100)])
img_layer0 = build_generator(h, sizes[0], sizes[1])
img_layer1 = build_generator(h, sizes[2], sizes[3])
img_layer2 = build_generator(h, sizes[4], sizes[5])
avgpool = AveragePooling2D(pool_size=(1, 8))
zero2one = avgpool(UpSampling2D(size=(4, 1))(img_layer0))
img_layer1 = inpainting_attention(img_layer1, zero2one)
one2two = AveragePooling2D(pool_size=(1, 2))(img_layer1)
img_layer2 = inpainting_attention(img_layer2, one2two)
outputs = [Activation('relu')(img_layer0),
Activation('relu')(img_layer1),
Activation('relu')(img_layer2)]
return outputs
# constrain w/ a tanh to dampen the unbounded nature of energy-space
mbd_energy = Activation('tanh')(minibatch_featurizer(K_energy))
# absolute deviation away from input energy. Technically we can learn
# this, but since we want to get as close as possible to conservation of
# energy, just coding it in is better
energy_well = Lambda(lambda x: K.abs(x[0] - x[1]))(
[total_energy, input_energy])
# binary y/n if it is over the input energy
well_too_big = Lambda(lambda x: 10 * K.cast(x > 5, K.floatx()))(
energy_well)
p = concatenate([
features,
scale(energies, 10),
scale(total_energy, 100), energy_well, well_too_big, mbd_energy
])
fake = Dense(1, activation='sigmoid', name='fakereal_output')(p)
discriminator_outputs = [fake, total_energy]
discriminator_losses = ['binary_crossentropy', 'mae']
# ACGAN case
if nb_classes > 1:
logger.info('running in ACGAN for discriminator mode since found {} '
'classes'.format(nb_classes))
aux = Dense(1, activation='sigmoid', name='auxiliary_output')(p)
discriminator_outputs.append(aux)
# change the loss depending on how many outputs on the auxiliary task
# showers to generate
input_folder = sys.argv[1]
output_folder = sys.argv[2]
epochs = int(sys.argv[3])
image_sets = int(sys.argv[4])
showers_to_generate = int(sys.argv[5])
latent_size = 1024
# input placeholders
latent = Input(shape=(latent_size, ), name='z') # noise
input_energy = Input(
shape=(1, ), dtype='float32') # requested energy of the particle shower
generator_inputs = [latent, input_energy]
# multiply the (scaled) energy into the latent space
h = Lambda(lambda x: x[0] * x[1])([latent, scale(input_energy, 100)])
# build three LAGAN-style generators (checkout out `build_generator` in architectures.py)
img_layer0 = build_generator(h, 3, 96)
img_layer1 = build_generator(h, 12, 12)
img_layer2 = build_generator(h, 12, 6)
# inpainting
# 0 --> 1
zero2one = AveragePooling2D(pool_size=(1,
8))(UpSampling2D(size=(4,
1))(img_layer0))
img_layer1 = inpainting_attention(img_layer1,
zero2one) # this function is in ops.py
# 1 --> 2
one2two = AveragePooling2D(pool_size=(1, 2))(img_layer1)
# constrain w/ a tanh to dampen the unbounded nature of energy-space
#mbd_energy = Activation('tanh')(minibatch_featurizer(K_energy))
# absolute deviation away from input energy. Technically we can learn
# this, but since we want to get as close as possible to conservation of
# energy, just coding it in is better
#energy_well = Lambda(
# lambda x: K.abs(x[0] - x[1])
#)([total_energy, input_energy])
# binary y/n if it is over the input energy
#well_too_big = Lambda(lambda x: 10 * K.cast(x > 5, K.floatx()))(energy_well)
p = concatenate([
features,
scale(energies, 10),
scale(total_energy, 100)
#energy_well,
#well_too_big,
#mbd_energy
])
fake = Dense(1, activation='sigmoid', name='fakereal_output')(p)
discriminator_outputs = [fake, total_energy]
discriminator_losses = ['binary_crossentropy', 'mae']
# ACGAN case
if nb_classes > 1:
logger.info('running in ACGAN for discriminator mode since found {} '
'classes'.format(nb_classes))
aux = Dense(1, activation='sigmoid', name='auxiliary_output')(p)