train=False),
batch_size=args.batch_size,
shuffle=True)
data_loaders['sup'] = train_loader_sup
data_loaders['unsup'] = train_loader_unsup
# how often would a supervised batch be encountered during inference
periodic_interval_batches = 5
# number of unsupervised examples
sup_num = len(data_loaders['sup']) * args.batch_size
unsup_num = len(data_loaders['unsup']) * args.batch_size
# setup the VAE
model = DIVA(args).to(device)
# setup the optimizer
optimizer = optim.Adam(model.parameters(), lr=args.lr)
# init
val_total_loss = []
val_class_err_d = []
val_class_err_y = []
best_loss = 1000.
best_y_acc = 0.
early_stopping_counter = 1
max_early_stopping = 100
torch.backends.cudnn.benchmark = False
np.random.seed(args.seed)
# Load supervised training
train_loader = data_utils.DataLoader(MnistRotated(args.list_train_domains,
args.list_test_domain,
args.num_supervised,
args.seed,
'./../dataset/',
train=True),
batch_size=args.batch_size,
shuffle=True,
**kwargs)
# setup the VAE
model = DIVA(args).to(device)
# setup the optimizer
optimizer = optim.Adam(model.parameters(), lr=args.lr)
best_loss = 1000.
best_y_acc = 0.
early_stopping_counter = 1
max_early_stopping = 100
# training loop
print('\nStart training:', args)
for epoch in range(1, args.epochs + 1):
model.beta_d = min(
[args.beta_d, args.beta_d * (epoch * 1.) / args.warmup])