def main():
try:
os.mkdir(args.plot_dir)
except:
pass
# settings
max_epoch = 1000
n_trains_per_epoch = 2
batchsize_positive = 100
batchsize_negative = 100
plotsize = 400
# config
config_energy_model = to_object(params_energy_model["config"])
config_generative_model = to_object(params_generative_model["config"])
# seed
np.random.seed(args.seed)
if args.gpu_enabled:
cuda.cupy.random.seed(args.seed)
# init weightnorm layers
if config_energy_model.use_weightnorm:
print "initializing weight normalization layers of the energy model ..."
x_positive = sampler.sample_from_swiss_roll(batchsize_positive * 10, 2,
10)
ddgm.compute_energy(x_positive)
if config_generative_model.use_weightnorm:
print "initializing weight normalization layers of the generative model ..."
x_negative = ddgm.generate_x(batchsize_negative * 10)
fixed_z = ddgm.sample_z(plotsize)
fixed_target = sampler.sample_from_swiss_roll(600, 2, 10)
progress = Progress()
for epoch in xrange(1, max_epoch):
progress.start_epoch(epoch, max_epoch)
sum_energy_positive = 0
sum_energy_negative = 0
sum_kld = 0
for t in xrange(n_trains_per_epoch):
# sample from data distribution
x_positive = sampler.sample_from_swiss_roll(
batchsize_positive, 2, 10)
# sample from generator
x_negative = ddgm.generate_x(batchsize_negative)
# train energy model
energy_positive = ddgm.compute_energy_sum(x_positive)
energy_negative = ddgm.compute_energy_sum(x_negative)
loss = energy_positive - energy_negative
ddgm.backprop_energy_model(loss)
# train generative model
x_negative = ddgm.generate_x(batchsize_negative)
kld = ddgm.compute_kld_between_generator_and_energy_model(
x_negative)
ddgm.backprop_generative_model(kld)
sum_energy_positive += float(energy_positive.data)
sum_energy_negative += float(energy_negative.data)
sum_kld += float(kld.data)
if t % 10 == 0:
progress.show(t, n_trains_per_epoch, {})
progress.show(
n_trains_per_epoch, n_trains_per_epoch, {
"x+": sum_energy_positive / n_trains_per_epoch,
"x-": sum_energy_negative / n_trains_per_epoch,
"KLD": int(sum_kld / n_trains_per_epoch)
})
ddgm.save(args.model_dir)
# init
fig = pylab.gcf()
fig.set_size_inches(8.0, 8.0)
pylab.clf()
plot(fixed_target, color="#bec3c7", s=20)
plot(ddgm.generate_x_from_z(fixed_z, as_numpy=True, test=True),
color="#e84c3d",
s=20)
# save
pylab.savefig("{}/{}.png".format(args.plot_dir, 100000 + epoch))
def main():
# load MNIST images
images = load_rgb_images(args.image_dir)
# config
config_energy_model = to_object(params_energy_model["config"])
config_generative_model = to_object(params_generative_model["config"])
# settings
max_epoch = 1000
n_trains_per_epoch = 500
batchsize_positive = 128
batchsize_negative = 128
plot_interval = 5
# seed
np.random.seed(args.seed)
if args.gpu_device != -1:
cuda.cupy.random.seed(args.seed)
# init weightnorm layers
if config_energy_model.use_weightnorm:
print "initializing weight normalization layers of the energy model ..."
x_positive = sample_from_data(images, batchsize_positive * 5)
ddgm.compute_energy(x_positive)
if config_generative_model.use_weightnorm:
print "initializing weight normalization layers of the generative model ..."
x_negative = ddgm.generate_x(batchsize_negative * 5)
progress = Progress()
for epoch in xrange(1, max_epoch):
progress.start_epoch(epoch, max_epoch)
sum_energy_positive = 0
sum_energy_negative = 0
sum_loss = 0
sum_kld = 0
for t in xrange(n_trains_per_epoch):
# sample from data distribution
x_positive = sample_from_data(images, batchsize_positive)
# sample from generator
x_negative = ddgm.generate_x(batchsize_negative)
# train energy model
energy_positive = ddgm.compute_energy_sum(x_positive)
energy_negative = ddgm.compute_energy_sum(x_negative)
loss = energy_positive - energy_negative
ddgm.backprop_energy_model(loss)
# train generative model
# TODO: KLD must be greater than or equal to 0
x_negative = ddgm.generate_x(batchsize_negative)
kld = ddgm.compute_kld_between_generator_and_energy_model(
x_negative)
ddgm.backprop_generative_model(kld)
sum_energy_positive += float(energy_positive.data)
sum_energy_negative += float(energy_negative.data)
sum_loss += float(loss.data)
sum_kld += float(kld.data)
progress.show(t, n_trains_per_epoch, {})
progress.show(
n_trains_per_epoch, n_trains_per_epoch, {
"x+": int(sum_energy_positive / n_trains_per_epoch),
"x-": int(sum_energy_negative / n_trains_per_epoch),
"loss": sum_loss / n_trains_per_epoch,
"kld": sum_kld / n_trains_per_epoch
})
ddgm.save(args.model_dir)
if epoch % plot_interval == 0 or epoch == 1:
plot(filename="epoch_{}_time_{}min".format(
epoch, progress.get_total_time()))
def main():
# settings
max_epoch = 1000
n_trains_per_epoch = 500
batchsize_positive = 100
batchsize_negative = 100
# config
config_energy_model = to_object(params_energy_model["config"])
config_generative_model = to_object(params_generative_model["config"])
# seed
np.random.seed(args.seed)
if args.gpu_enabled:
cuda.cupy.random.seed(args.seed)
# init weightnorm layers
if config_energy_model.use_weightnorm:
print "initializing weight normalization layers of the energy model ..."
x_positive = sampler.sample_from_gaussian_mixture(
batchsize_positive * 10, 2, 10)
ddgm.compute_energy(x_positive)
if config_generative_model.use_weightnorm:
print "initializing weight normalization layers of the generative model ..."
x_negative = ddgm.generate_x(batchsize_negative * 10)
progress = Progress()
for epoch in xrange(1, max_epoch):
progress.start_epoch(epoch, max_epoch)
sum_energy_positive = 0
sum_energy_negative = 0
sum_kld = 0
for t in xrange(n_trains_per_epoch):
# sample from data distribution
x_positive = sampler.sample_from_gaussian_mixture(
batchsize_positive, 2, 10)
# sample from generator
x_negative = ddgm.generate_x(batchsize_negative)
# train energy model
energy_positive = ddgm.compute_energy_sum(x_positive)
energy_negative = ddgm.compute_energy_sum(x_negative)
loss = energy_positive - energy_negative
ddgm.backprop_energy_model(loss)
# train generative model
x_negative = ddgm.generate_x(batchsize_negative)
kld = ddgm.compute_kld_between_generator_and_energy_model(
x_negative)
ddgm.backprop_generative_model(kld)
sum_energy_positive += float(energy_positive.data)
sum_energy_negative += float(energy_negative.data)
sum_kld += float(kld.data)
if t % 10 == 0:
progress.show(t, n_trains_per_epoch, {})
progress.show(
n_trains_per_epoch, n_trains_per_epoch, {
"x+": sum_energy_positive / n_trains_per_epoch,
"x-": sum_energy_negative / n_trains_per_epoch,
"KLD": int(sum_kld / n_trains_per_epoch)
})
ddgm.save(args.model_dir)