def train_epoch_dense(model, optimizer, device, data_loader, epoch,
batch_size):
model.train()
epoch_loss = 0
epoch_train_f1 = 0
nb_data = 0
gpu_mem = 0
optimizer.zero_grad()
for iter, (x_no_edge_feat, x_with_edge_feat, labels,
edge_list) in enumerate(data_loader):
if x_no_edge_feat is not None:
x_no_edge_feat = x_no_edge_feat.to(device)
if x_with_edge_feat is not None:
x_with_edge_feat = x_with_edge_feat.to(device)
labels = labels.to(device)
edge_list = edge_list[0].to(device), edge_list[1].to(device)
scores = model.forward(x_no_edge_feat, x_with_edge_feat, edge_list)
loss = model.loss(scores, labels)
loss.backward()
if not (iter % batch_size):
optimizer.step()
optimizer.zero_grad()
epoch_loss += loss.detach().item()
epoch_train_f1 += binary_f1_score(scores, labels)
epoch_loss /= (iter + 1)
epoch_train_f1 /= (iter + 1)
return epoch_loss, epoch_train_f1, optimizer
def train_epoch(model, optimizer, device, data_loader, epoch):
model.train()
epoch_loss = 0
epoch_train_f1 = 0
nb_data = 0
gpu_mem = 0
for iter, (batch_graphs, batch_labels, batch_snorm_n,
batch_snorm_e) in enumerate(data_loader):
batch_x = batch_graphs.ndata['feat'].to(device) # num x feat
batch_e = batch_graphs.edata['feat'].to(device)
batch_labels = batch_labels.to(device)
batch_snorm_e = batch_snorm_e.to(device)
batch_snorm_n = batch_snorm_n.to(device) # num x 1
optimizer.zero_grad()
batch_scores = model.forward(batch_graphs, batch_x, batch_e,
batch_snorm_n, batch_snorm_e)
loss = model.loss(batch_scores, batch_labels)
loss.backward()
optimizer.step()
epoch_loss += loss.detach().item()
epoch_train_f1 += binary_f1_score(batch_scores, batch_labels)
epoch_loss /= (iter + 1)
epoch_train_f1 /= (iter + 1)
return epoch_loss, epoch_train_f1, optimizer
def evaluate_network_sparse(model, device, data_loader, epoch):
model.eval()
epoch_test_loss = 0
epoch_test_f1 = 0
nb_data = 0
with torch.no_grad():
for iter, (batch_graphs, batch_labels) in enumerate(data_loader):
batch_x = batch_graphs.ndata['feat'].to(device)
batch_e = batch_graphs.edata['feat'].to(device)
batch_labels = batch_labels.to(device)
batch_scores = model.forward(batch_graphs, batch_x, batch_e)
loss = model.loss(batch_scores, batch_labels)
epoch_test_loss += loss.detach().item()
epoch_test_f1 += binary_f1_score(batch_scores, batch_labels)
epoch_test_loss /= (iter + 1)
epoch_test_f1 /= (iter + 1)
return epoch_test_loss, epoch_test_f1
def train_epoch(model, optimizer, device, data_loader, epoch):
model.train()
epoch_loss = 0
epoch_train_f1 = 0
nb_data = 0
print(
time.strftime("%Y-%m-%d %H:%M:%S", time.localtime()) + ' | ' +
'Epoch [{:>2d}]: learning rate: [{:.10f}]'.format(
epoch + 1, optimizer.param_groups[0]['lr']))
for iter, (batch_graphs, batch_labels, batch_snorm_n,
batch_snorm_e) in enumerate(data_loader):
batch_x = batch_graphs.ndata['feat'].to(device) # num x feat
batch_e = batch_graphs.edata['feat'].to(device)
batch_labels = batch_labels.to(device)
batch_snorm_e = batch_snorm_e.to(device)
batch_snorm_n = batch_snorm_n.to(device) # num x 1
optimizer.zero_grad()
batch_scores = model.forward(batch_graphs, batch_x, batch_e,
batch_snorm_n, batch_snorm_e)
loss = model.loss(batch_scores, batch_labels)
loss.backward()
optimizer.step()
epoch_loss += loss.detach().item()
epoch_train_f1 += binary_f1_score(batch_scores, batch_labels)
if iter % 10 == 0:
print(
time.strftime("%Y-%m-%d %H:%M:%S", time.localtime()) + ' | ' +
'Iter [{:>3d}]: Loss [{:.4f}] Train F1 [{:.4f}]'.format(
iter, loss, epoch_train_f1 / (iter + 1)))
print(
time.strftime("%Y-%m-%d %H:%M:%S", time.localtime()) + ' | ' +
'Iter [{:>3d}]: Loss [{:.4f}] Train F1 [{:.4f}]'.format(
iter, loss, epoch_train_f1 / (iter + 1)))
epoch_loss /= (iter + 1)
epoch_train_f1 /= (iter + 1)
return epoch_loss, epoch_train_f1, optimizer
def evaluate_network_dense(model, device, data_loader, epoch):
model.eval()
epoch_test_loss = 0
epoch_test_f1 = 0
nb_data = 0
with torch.no_grad():
for iter, (x_no_edge_feat, x_with_edge_feat, labels, edge_list) in enumerate(data_loader):
if x_no_edge_feat is not None:
x_no_edge_feat = x_no_edge_feat.to(device)
if x_with_edge_feat is not None:
x_with_edge_feat = x_with_edge_feat.to(device)
labels = labels.to(device)
edge_list = edge_list[0].to(device), edge_list[1].to(device)
scores = model.forward(x_no_edge_feat, x_with_edge_feat, edge_list)
loss = model.loss(scores, labels)
epoch_test_loss += loss.detach().item()
epoch_test_f1 += binary_f1_score(scores, labels)
epoch_test_loss /= (iter + 1)
epoch_test_f1 /= (iter + 1)
return epoch_test_loss, epoch_test_f1
