if mask == "val": offset = 3
else: offset = 5
z = model.lp_encode(data[0])
link_logits = model.lp_decode(z, torch.stack(data[offset].edges()),
torch.stack(data[offset + 1].edges()))
link_probs = link_logits.sigmoid()
link_labels = get_link_labels(torch.stack(data[offset].edges()),
torch.stack(data[offset + 1].edges()))
result = roc_auc_score(link_labels.cpu().numpy(), link_probs.cpu().numpy())
return result
res = []
model_hp, _ = get_encoder_decoder_hp(args.model, decoupled=False)
for seed in tqdm(range(1234, 1234 + args.repeat)):
setup_seed(seed)
g = dataset[0]
splitted = list(split_train_valid_test(g))
if args.model == 'gcn' or args.model == 'gat':
splitted[0] = dgl.add_self_loop(splitted[0])
splitted = [g.to(device) for g in splitted]
if args.model == 'gcn':
model = AutoGCN(
input_dimension=splitted[0].ndata['feat'].shape[1],
output_dimension=2,
parser.add_argument('--epoch', type=int, default=200)
args = parser.parse_args()
# seed = 100
dataset = build_dataset_from_name(args.dataset.lower())
dataset = to_dgl_dataset(dataset)
data = dataset[0].to(args.device)
num_features = data.ndata['feat'].size(1)
num_classes = data.ndata['label'].max().item() + 1
label = data.ndata['label']
test_mask = data.ndata['test_mask']
accs = []
model_hp, decoder_hp = get_encoder_decoder_hp(args.model)
for seed in tqdm(range(args.repeat)):
set_seed(seed)
trainer = NodeClassificationFullTrainer(
model=args.model,
num_features=num_features,
num_classes=num_classes,
device=args.device,
init=False,
feval=['acc'],
loss="nll_loss",
).duplicate_from_hyper_parameter({
"trainer": {
"max_epoch": args.epoch,
parser.add_argument('--repeat', type=int, default=50)
parser.add_argument('--model', type=str, choices=['gat', 'gcn', 'sage'], default='gat')
parser.add_argument('--lr', type=float, default=0.01)
parser.add_argument('--weight_decay', type=float, default=0.0)
parser.add_argument('--epoch', type=int, default=200)
parser.add_argument('--debug', action='store_true', default=False)
args = parser.parse_args()
# seed = 100
dataset = Planetoid(os.path.expanduser('~/.cache-autogl'), args.dataset, transform=T.NormalizeFeatures())
data = dataset[0].to(args.device)
accs = []
model_hp, decoder_hp = get_encoder_decoder_hp(args.model, decoupled=True)
for seed in tqdm(range(args.repeat)):
set_seed(seed)
if args.model == 'gat':
model = GATEncoderMaintainer(
input_dimension=dataset.num_node_features,
final_dimension=dataset.num_classes,
device=args.device
).from_hyper_parameter(model_hp)
elif args.model == 'gcn':
model = GCNEncoderMaintainer(
input_dimension=dataset.num_node_features,
final_dimension=dataset.num_classes,
device=args.device
