class_loss_f, nb_classes)
if train:
resume = args['resume']
num_epoch = int(args['epoch'])
sampler = SampleImage(10, 10, code_size)
if resume == 'model':
gan_model.load_checkpoint(checkpoint_path, all=False)
elif resume == 'checkpoint':
gan_model.load_checkpoint(checkpoint_path, all=True)
trainer = Trainer(trick_dict={
'label_smooth': {
'valid_range': 0.95,
'fake_range': 0
}
})
data_loader = get_mnist_data_loader(train=True)
data_loader_label = get_mnist_subset_data_loader(train=True,
fraction=fraction)
trainer.train(num_epoch, data_loader, data_loader_label, gan_model, 1,
checkpoint_path, 5, [sampler])
else:
# evaluate accuracy on test set
gan_model.load_checkpoint(checkpoint_path, all=False)
data_loader = get_mnist_data_loader(train=False)
acc = compute_accuracy(data_loader, gan_model)
print('The test accuracy of discriminator is {:.4f}'.format(acc))
code_size = 10
checkpoint_path = './checkpoint/vae_mnist.ckpt'
learning_rate = 1e-3
generator = Encoder(code_size)
decoder = Decoder(code_size)
prior = Normal(loc=torch.zeros(size=(code_size, )),
scale=torch.ones(size=(code_size, )))
optimizer = torch.optim.Adam(
chain(generator.parameters(), decoder.parameters()), learning_rate)
model = VAE(generator, decoder, prior, optimizer)
sampler = SampleImage(10, 10)
data_loader = get_mnist_subset_data_loader(train=True,
transform=transform,
fraction=100)
reconstruct = Reconstruction(next(iter(data_loader))[0])
summary_writer = SummaryWriter('runs/vae_mnist')
if train:
visualize_data_loader = get_mnist_subset_data_loader(
train=True, transform=transform, fraction=1000)
visualize_callback = VisualizeLatent(visualize_data_loader,
method='pca')
resume = args['resume']
num_epoch = int(args['epoch'])
if resume == 'model':
model.load_checkpoint(checkpoint_path, all=False)
generator = Generator(code_size=code_size,
weight_init=weights_init_normal,
weight_norm=weight_norm)
encoder = Encoder(code_size=code_size,
weight_init=weights_init_normal,
weight_norm=weight_norm)
optimizer_G = torch.optim.Adam(chain(generator.parameters(),
encoder.parameters()),
lr_G,
betas=(0.5, 0.999))
gan_model = ALI(generator, encoder, discriminator, optimizer_G,
optimizer_D, code_size, validity_loss_f)
mnist_subset = get_mnist_subset_data_loader(train=True, fraction=100)
sampler = SampleImage(10, code_size)
reconstruction = Reconstruction(mnist_subset)
visualize_latent = VisualizeLatent(mnist_subset, method='pca')
if train:
if resume == 'model':
gan_model.load_checkpoint(checkpoint_path, all=False)
elif resume == 'checkpoint':
gan_model.load_checkpoint(checkpoint_path, all=True)
trainer = Trainer(trick_dict={
'label_smooth': {
'valid_range': 0.95,
betas=(0.5, 0.999))
gaussian_sampler = GaussianSampler()
gan_model = AAE(generator,
decoder,
discriminator,
optimizer_G,
optimizer_D,
code_size,
validity_loss_f,
recon_loss_f,
gaussian_sampler,
alpha=alpha)
recon_data_loader = get_mnist_subset_data_loader(train=True, fraction=100)
recon_callback = Reconstruction(recon_data_loader)
sampler = SampleImage(10, 10)
if train:
visualize_data_loader = get_mnist_subset_data_loader(train=True,
fraction=100)
visualize_callback = VisualizeLatent(visualize_data_loader,
method='pca')
resume = args['resume']
num_epoch = int(args['epoch'])
if resume == 'model':
gan_model.load_checkpoint(checkpoint_path, all=False)
elif resume == 'checkpoint':