def noise_sampler(bs):
return np.random.normal(0.0, 1.0, [bs, 3])
if __name__ == '__main__':
from objectives.bayes_logistic_regression.synthetic import Synthetic
from models.discriminator import MLPDiscriminator
from models.generator import create_nice_network
from train.wgan_nll import Trainer
os.environ['CUDA_VISIBLE_DEVICES'] = '0'
energy_fn = Synthetic(batch_size=32)
discriminator = MLPDiscriminator([800, 800, 800])
generator = create_nice_network(3, 10, [
([400], 'v1', False),
([400, 400], 'x1', True),
([400], 'v2', False),
])
trainer = Trainer(generator,
energy_fn,
discriminator,
noise_sampler,
b=16,
m=4,
eta=5)
trainer.train(bootstrap_steps=3000, bootstrap_burn_in=1000)
def noise_sampler(bs):
return np.random.normal(0.0, 1.0, [bs, 25])
if __name__ == '__main__':
from objectives.bayes_logistic_regression.german import German
from models.discriminator import MLPDiscriminator
from models.generator import create_nice_network
from train.wgan_nll import Trainer
os.environ['CUDA_VISIBLE_DEVICES'] = '0'
energy_fn = German(batch_size=32)
discriminator = MLPDiscriminator([800, 800, 800])
generator = create_nice_network(25, 50, [
([400], 'v1', False),
([400, 400], 'x1', True),
([400], 'v2', False),
])
trainer = Trainer(generator,
energy_fn,
discriminator,
noise_sampler,
b=16,
m=2)
trainer.train(bootstrap_steps=5000,
bootstrap_burn_in=1000,
bootstrap_discard_ratio=0.5)
import os
import sys
sys.path.append(os.getcwd())
def noise_sampler(bs):
return np.random.normal(0.0, 1.0, [bs, 2])
if __name__ == '__main__':
from objectives.expression.mog6 import MixtureOfGaussians
from models.discriminator import MLPDiscriminator
from models.generator import create_nice_network
from train.wgan_nll import Trainer
os.environ['CUDA_VISIBLE_DEVICES'] = ''
energy_fn = MixtureOfGaussians(display=False)
discriminator = MLPDiscriminator([400, 400, 400])
generator = create_nice_network(
2, 2,
[
([400], 'v1', False),
([400], 'x1', True),
([400], 'v2', False),
]
)
trainer = Trainer(generator, energy_fn, discriminator, noise_sampler, b=8, m=2)
trainer.train()