cut_idx) #temp_model 用的是当前model后半部分，models[]前半部分
for i in range(task + 1): #循环不同task
acc = test_model(temp_model, labels[i], test_loader[i],
gpu)
print('Device Task: {}\tTest Task: {}\tAccuracy: {}'.
format(sub_task, i, acc))
with open(result_file, 'a') as f:
# Current server parameter (task) + device parameter (task) --> training a given task
f.write('{}\t{}\t{}\t{}\t{}\t{}\n'.format(
task, iteration, loss, sub_task, i, acc))
ewcs.append(splitEWC(model, train_loader[task], cut_idx, gpu))
if __name__ == '__main__':
if args.dataset == 'permuted':
train_loader, test_loader, labels = get_permute_mnist(
args.bsz, args.num_task, [1829, 241, 43, 12649, 4443])
elif args.dataset == 'rotated':
train_loader, test_loader, labels = get_rotate_mnist(
args.bsz, args.num_task)
else:
train_loader, test_loader, labels = get_split_mnist(
args.bsz, args.num_task)
if args.method == 'ewc':
ewc_process(train_loader,
test_loader,
labels,
result_file='./result/{}_{}_{}.txt'.format(
args.method, args.dataset, args.split))
elif args.method == 'online':
ewc_process(train_loader,
print('Device Task: {}\tTest Task: {}\tAccuracy: {}'.
format(sub_task, i, acc))
with open(result_file, 'a') as f:
# Current server parameter (task) + device parameter (task) --> training a given task
f.write('{}\t{}\t{}\t{}\t{}\t{}\n'.format(
task, iteration, loss, sub_task, i, acc))
if iteration % 20 == 0:
for param_group in optimizer.param_groups:
param_group['lr'] *= 0.95
ewcs.append(splitEWC(model, train_loader[task], cut_idx, gpu))
if __name__ == '__main__':
if args.dataset == 'permuted':
# train_loader, test_loader, labels = get_permute_mnist(args.bsz, args.num_task, [1829, 241, 43, 12649, 4443])
train_loader, test_loader, labels = get_permute_mnist(
args.bsz, args.num_task)
elif args.dataset == 'rotated':
train_loader, test_loader, labels = get_rotate_mnist(
args.bsz, args.num_task)
elif 'mnist' in args.dataset:
train_loader, test_loader, labels = get_split_mnist(
args.bsz, args.num_task)
elif 'cifar' in args.dataset:
train_loader, test_loader, labels = get_split_cifar(
args.bsz, args.num_task)
if args.method == 'ewc':
ewc_process(train_loader,
test_loader,
labels,
'class' in args.dataset,