elif args.net == "mobilenet":
G = torchvision.models.mobilenet_v2(pretrained=True)
inc = 1280
G.classifier = torch.nn.Identity()
else:
raise ValueError('Model cannot be recognized.')
if "resnet" in args.net or "densenet" in args.net or "mobilenet" in args.net:
F1 = Predictor_deep(num_class=len(class_list), inc=inc)
else:
F1 = Predictor(num_class=len(class_list), inc=inc, temp=args.T)
device = torch.device("cuda:" + args.device)
G = G.to(device)
F1 = F1.to(device)
G.load_state_dict(
torch.load(
os.path.join(
args.checkpath,
"G_iter_model_{}_{}_{}_to_{}_num_{}.pth.tar".format(
args.method, args.net, args.source, args.target, args.num))))
F1.load_state_dict(
torch.load(
os.path.join(
args.checkpath,
"F1_iter_model_{}_{}_{}_to_{}_num_{}.pth.tar".format(
args.method, args.net, args.source, args.target, args.num))))
im_data_t = torch.FloatTensor(1)
params += [{
'params': [value],
'lr': args.multi * 10,
'weight_decay': 0.0005
}]
if "resnet" in args.net or "densenet" in args.net or "mobilenet" in args.net or "inception" in args.net:
F1 = Predictor_deep(num_class=len(class_list), inc=inc)
else:
F1 = Predictor(num_class=len(class_list), inc=inc, temp=args.T)
weights_init(F1)
lr = args.lr
device = torch.device("cuda:" + args.device)
G.to(device)
F1.to(device)
im_data_s = torch.FloatTensor(1)
im_data_t = torch.FloatTensor(1)
im_data_tu = torch.FloatTensor(1)
gt_labels_s = torch.LongTensor(1)
gt_labels_t = torch.LongTensor(1)
sample_labels_t = torch.LongTensor(1)
sample_labels_s = torch.LongTensor(1)
im_data_s = im_data_s.to(device)
im_data_t = im_data_t.to(device)
im_data_tu = im_data_tu.to(device)
gt_labels_s = gt_labels_s.to(device)
gt_labels_t = gt_labels_t.to(device)
sample_labels_t = sample_labels_t.to(device)
