ac_gan = ACGAN(100, 10, 1, device)
print(ac_gan)
dataset = datasets.MNIST(root='./data',
download=True,
transform=transforms.Compose([
transforms.Resize(64),
transforms.ToTensor(),
transforms.Normalize((0.5, ), (0.5, ))
]))
dataloader = torch.utils.data.DataLoader(dataset,
batch_size=100,
shuffle=True,
num_workers=2)
src_criterion = nn.MSELoss()
c_criterion = nn.CrossEntropyLoss()
d_optim = optim.Adam(ac_gan.discriminator.parameters(),
lr=0.0002,
betas=(0.5, 0.999))
g_optim = optim.Adam(ac_gan.generator.parameters(),
lr=0.0002,
betas=(0.5, 0.999))
ac_gan.train(dataloader, 50, src_criterion, c_criterion, g_optim, d_optim)
torch.save(ac_gan.generator.state_dict(), './mnist-acgan-g.pth')
torch.save(ac_gan.discriminator.state_dict(), './mnist-acgan-d.pth')
month_label_train = month_label[:num_train]
month_label_val = month_label[num_train:]
day_label_train = day_label[:num_train]
day_label_val = day_label[num_train:]
print(x_train.shape, x_val.shape, month_label_train.shape,
month_label_val.shape, day_label_train.shape,
day_label_val.shape)
weight_path = 'weights/pecan' + '_user_' + str(user_id) + '_'
model = Model(input_dim=2, window_length=96, weight_path=weight_path)
if train:
num_epoch = args['num_epoch']
print(model.discriminator.summary())
print(model.generator.summary())
model.train([x_train, month_label_train, day_label_train],
[x_val, month_label_val, day_label_val],
num_epoch=num_epoch)
else:
x_generated = model.generate_by_date(1461)
usage_generated = x_generated[:, :, 0]
gen_generated = x_generated[:, :, 1]
usage_generated_recover = usage_recover(usage_generated)
gen_generated_recover = gen_recover(gen_generated)
data = np.stack((usage_generated_recover, gen_generated_recover),
axis=-1)
print('The shape of the synthetic data is {}'.format(data.shape))
np.savez_compressed('synthetic/user_' + str(user_id) +
'_acgan.npz',
data=data)