def gat(*args, **kwargs):
return GAT(*args, **kwargs)
feature_size = [1, 2708, 1433]
biases_size = [1, 2708, 2708]
num_classes = 7
hid_units = GatConfig.hid_units
n_heads = GatConfig.n_heads
feature = np.random.uniform(0.0, 1.0, size=feature_size).astype(np.float32)
biases = np.random.uniform(0.0, 1.0, size=biases_size).astype(np.float64)
feature_size = feature.shape[2]
num_nodes = feature.shape[1]
gat_net = GAT(feature_size,
num_classes,
num_nodes,
hid_units,
n_heads,
attn_drop=0.0,
ftr_drop=0.0)
gat_net.set_train(False)
load_checkpoint(args.ckpt_file, net=gat_net)
gat_net.add_flags_recursive(fp16=True)
export(gat_net,
Tensor(feature),
Tensor(biases),
file_name=args.file_name,
file_format=args.file_format)
def train():
"""Train GAT model."""
parser = argparse.ArgumentParser()
parser.add_argument('--data_dir',
type=str,
default='./data/cora/cora_mr',
help='Data dir')
parser.add_argument('--train_nodes_num',
type=int,
default=140,
help='Nodes numbers for training')
parser.add_argument('--eval_nodes_num',
type=int,
default=500,
help='Nodes numbers for evaluation')
parser.add_argument('--test_nodes_num',
type=int,
default=1000,
help='Nodes numbers for test')
args = parser.parse_args()
if not os.path.exists("ckpts"):
os.mkdir("ckpts")
context.set_context(mode=context.GRAPH_MODE,
device_target="Ascend",
save_graphs=False)
# train parameters
hid_units = GatConfig.hid_units
n_heads = GatConfig.n_heads
early_stopping = GatConfig.early_stopping
lr = GatConfig.lr
l2_coeff = GatConfig.l2_coeff
num_epochs = GatConfig.num_epochs
feature, biases, y_train, train_mask, y_val, eval_mask, y_test, test_mask = load_and_process(
args.data_dir, args.train_nodes_num, args.eval_nodes_num,
args.test_nodes_num)
feature_size = feature.shape[2]
num_nodes = feature.shape[1]
num_class = y_train.shape[2]
gat_net = GAT(feature,
biases,
feature_size,
num_class,
num_nodes,
hid_units,
n_heads,
attn_drop=GatConfig.attn_dropout,
ftr_drop=GatConfig.feature_dropout)
gat_net.add_flags_recursive(fp16=True)
eval_net = LossAccuracyWrapper(gat_net, num_class, y_val, eval_mask,
l2_coeff)
train_net = TrainGAT(gat_net, num_class, y_train, train_mask, lr, l2_coeff)
train_net.set_train(True)
val_acc_max = 0.0
val_loss_min = np.inf
for _epoch in range(num_epochs):
train_result = train_net()
train_loss = train_result[0].asnumpy()
train_acc = train_result[1].asnumpy()
eval_result = eval_net()
eval_loss = eval_result[0].asnumpy()
eval_acc = eval_result[1].asnumpy()
print(
"Epoch:{}, train loss={:.5f}, train acc={:.5f} | val loss={:.5f}, val acc={:.5f}"
.format(_epoch, train_loss, train_acc, eval_loss, eval_acc))
if eval_acc >= val_acc_max or eval_loss < val_loss_min:
if eval_acc >= val_acc_max and eval_loss < val_loss_min:
val_acc_model = eval_acc
val_loss_model = eval_loss
_exec_save_checkpoint(train_net.network, "ckpts/gat.ckpt")
val_acc_max = np.max((val_acc_max, eval_acc))
val_loss_min = np.min((val_loss_min, eval_loss))
curr_step = 0
else:
curr_step += 1
if curr_step == early_stopping:
print("Early Stop Triggered!, Min loss: {}, Max accuracy: {}".
format(val_loss_min, val_acc_max))
print(
"Early stop model validation loss: {}, accuracy{}".format(
val_loss_model, val_acc_model))
break
gat_net_test = GAT(feature,
biases,
feature_size,
num_class,
num_nodes,
hid_units,
n_heads,
attn_drop=0.0,
ftr_drop=0.0)
load_checkpoint("ckpts/gat.ckpt", net=gat_net_test)
gat_net_test.add_flags_recursive(fp16=True)
test_net = LossAccuracyWrapper(gat_net_test, num_class, y_test, test_mask,
l2_coeff)
test_result = test_net()
print("Test loss={}, test acc={}".format(test_result[0], test_result[1]))
def train(args_opt):
"""Train GAT model."""
if not os.path.exists("ckpts"):
os.mkdir("ckpts")
# train parameters
hid_units = GatConfig.hid_units
n_heads = GatConfig.n_heads
early_stopping = GatConfig.early_stopping
lr = GatConfig.lr
l2_coeff = GatConfig.l2_coeff
num_epochs = GatConfig.num_epochs
feature, biases, y_train, train_mask, y_val, eval_mask, y_test, test_mask = load_and_process(
args_opt.data_dir, args_opt.train_nodes_num, args_opt.eval_nodes_num,
args_opt.test_nodes_num)
feature_size = feature.shape[2]
num_nodes = feature.shape[1]
num_class = y_train.shape[2]
gat_net = GAT(feature,
biases,
feature_size,
num_class,
num_nodes,
hid_units,
n_heads,
attn_drop=GatConfig.attn_dropout,
ftr_drop=GatConfig.feature_dropout)
gat_net.add_flags_recursive(fp16=True)
eval_net = LossAccuracyWrapper(gat_net, num_class, y_val, eval_mask,
l2_coeff)
train_net = TrainGAT(gat_net, num_class, y_train, train_mask, lr, l2_coeff)
train_net.set_train(True)
val_acc_max = 0.0
val_loss_min = np.inf
for _epoch in range(num_epochs):
train_result = train_net()
train_loss = train_result[0].asnumpy()
train_acc = train_result[1].asnumpy()
eval_result = eval_net()
eval_loss = eval_result[0].asnumpy()
eval_acc = eval_result[1].asnumpy()
print(
"Epoch:{}, train loss={:.5f}, train acc={:.5f} | val loss={:.5f}, val acc={:.5f}"
.format(_epoch, train_loss, train_acc, eval_loss, eval_acc))
if eval_acc >= val_acc_max or eval_loss < val_loss_min:
if eval_acc >= val_acc_max and eval_loss < val_loss_min:
val_acc_model = eval_acc
val_loss_model = eval_loss
save_checkpoint(train_net.network, "ckpts/gat.ckpt")
val_acc_max = np.max((val_acc_max, eval_acc))
val_loss_min = np.min((val_loss_min, eval_loss))
curr_step = 0
else:
curr_step += 1
if curr_step == early_stopping:
print("Early Stop Triggered!, Min loss: {}, Max accuracy: {}".
format(val_loss_min, val_acc_max))
print(
"Early stop model validation loss: {}, accuracy{}".format(
val_loss_model, val_acc_model))
break
gat_net_test = GAT(feature,
biases,
feature_size,
num_class,
num_nodes,
hid_units,
n_heads,
attn_drop=0.0,
ftr_drop=0.0)
load_checkpoint("ckpts/gat.ckpt", net=gat_net_test)
gat_net_test.add_flags_recursive(fp16=True)
test_net = LossAccuracyWrapper(gat_net_test, num_class, y_test, test_mask,
l2_coeff)
test_result = test_net()
print("Test loss={}, test acc={}".format(test_result[0], test_result[1]))