def main(args):
save_folder = '%s_%s' % (args.dataset, args.affix)
log_folder = os.path.join(args.log_root, save_folder)
model_folder = os.path.join(args.model_root, save_folder)
makedirs(log_folder)
makedirs(model_folder)
setattr(args, 'log_folder', log_folder)
setattr(args, 'model_folder', model_folder)
logger = create_logger(log_folder, args.todo, 'info')
print_args(args, logger)
model = Model(i_c=1, n_c=10)
attack = FastGradientSignUntargeted(model,
args.epsilon,
args.alpha,
min_val=0,
max_val=1,
max_iters=args.k,
_type=args.perturbation_type)
if torch.cuda.is_available():
model.cuda()
trainer = Trainer(args, logger, attack)
if args.todo == 'train':
tr_dataset = tv.datasets.MNIST(args.data_root,
train=True,
transform=tv.transforms.ToTensor(),
download=True)
tr_loader = DataLoader(tr_dataset,
batch_size=args.batch_size,
shuffle=True,
num_workers=4)
# evaluation during training
te_dataset = tv.datasets.MNIST(args.data_root,
train=False,
transform=tv.transforms.ToTensor(),
download=True)
te_loader = DataLoader(te_dataset,
batch_size=args.batch_size,
shuffle=False,
num_workers=4)
trainer.train(model, tr_loader, te_loader, args.adv_train)
elif args.todo == 'test':
pass
else:
raise NotImplementedError
def main(args):
if args.todo == 'train' or args.todo == 'valid':
folder_name = '%dway_%dshot_%s_%s' % (args.nway, args.shots,
args.model_type, args.affix)
model_folder = os.path.join(args.model_root, folder_name)
log_folder = os.path.join(args.log_root, folder_name)
mkdir(args.model_root)
mkdir(args.log_root)
mkdir(model_folder)
mkdir(log_folder)
setattr(args, 'model_folder', model_folder)
setattr(args, 'log_folder', log_folder)
logger = create_logger(log_folder, args.todo)
print_args(args, logger)
tr_dataloader = self_DataLoader(args.data_root,
train=True,
dataset=args.dataset,
seed=args.seed,
nway=args.nway)
trainer_dict = {
'args': args,
'logger': logger,
'tr_dataloader': tr_dataloader
}
trainer = Trainer(trainer_dict)
###########################################
## pretrain CNN embedding
if args.pretrain:
if args.pretrain_dir != '':
pretrain_path = os.path.join(args.pretrain_dir,
'pretrain_model.pth')
trainer.load_pretrain(pretrain_path)
else:
pretr_tr_data, pretr_tr_label = tr_dataloader.get_full_data_list(
) # already shuffled the data
va_size = int(0.1 * len(pretr_tr_data))
pretr_tr_dataset = self_Dataset(pretr_tr_data[va_size:],
pretr_tr_label[va_size:])
pretr_va_dataset = self_Dataset(pretr_tr_data[:va_size],
pretr_tr_label[:va_size])
logger.info('start pretraining...')
trainer.pretrain(pretr_tr_dataset, pretr_va_dataset)
logger.info('finish pretraining...')
###########################################
## load the model trained before
if args.load:
model_path = os.path.join(args.load_dir, 'model.pth')
trainer.load_model(model_path)
###########################################
## start training
trainer.train()
elif args.todo == 'test':
print(args.load_dir)
logger = create_logger('', args.todo)
print_args(args, logger)
te_dataloader = self_DataLoader(args.data_root,
train=False,
dataset=args.dataset,
seed=args.seed,
nway=args.nway)
trainer_dict = {'args': args, 'logger': logger}
trainer = Trainer(trainer_dict)
test_data_list, test_label_list = te_dataloader.get_few_data_list()
test_data_array, test_label_array = np.stack(
test_data_list), np.hstack(test_label_list)
if args.load:
model_path = os.path.join(args.load_dir, 'model.pth')
trainer.load_model(model_path)
test_pred = trainer.test(test_data_array, te_dataloader)
print(test_pred.shape, test_label_array.shape)
correct = (test_pred == test_label_array).sum()
test_acc = (test_pred == test_label_array).mean() * 100.0
print('test_acc: %.4f %%, correct: %d / %d' %
(test_acc, correct, len(test_label_array)))
def main(args):
save_folder = '%s_%s' % (args.dataset, args.affix)
log_folder = os.path.join(args.log_root, save_folder)
model_folder = os.path.join(args.model_root, save_folder)
makedirs(log_folder)
makedirs(model_folder)
setattr(args, 'log_folder', log_folder)
setattr(args, 'model_folder', model_folder)
logger = create_logger(log_folder, args.todo, 'info')
print_args(args, logger)
model = WideResNet(depth=34, num_classes=10, widen_factor=10, dropRate=0.0)
attack = FastGradientSignUntargeted(model,
args.epsilon,
args.alpha,
min_val=0,
max_val=1,
max_iters=args.k,
_type=args.perturbation_type)
if torch.cuda.is_available():
model.cuda()
trainer = Trainer(args, logger, attack)
if args.todo == 'train':
transform_train = tv.transforms.Compose([
tv.transforms.ToTensor(),
tv.transforms.Lambda(lambda x: F.pad(x.unsqueeze(0),
(4,4,4,4), mode='constant', value=0).squeeze()),
tv.transforms.ToPILImage(),
tv.transforms.RandomCrop(32),
tv.transforms.RandomHorizontalFlip(),
tv.transforms.ToTensor(),
])
tr_dataset = tv.datasets.CIFAR10(args.data_root,
train=True,
transform=transform_train,
download=True)
tr_loader = DataLoader(tr_dataset, batch_size=args.batch_size, shuffle=True, num_workers=4)
# evaluation during training
te_dataset = tv.datasets.CIFAR10(args.data_root,
train=False,
transform=tv.transforms.ToTensor(),
download=True)
te_loader = DataLoader(te_dataset, batch_size=args.batch_size, shuffle=False, num_workers=4)
trainer.train(model, tr_loader, te_loader, args.adv_train)
elif args.todo == 'test':
pass
else:
raise NotImplementedError
import models
from models.projector import Projector
from argument import linear_parser, print_args
from utils import progress_bar, checkpoint
from collections import OrderedDict
from attack_lib import FastGradientSignUntargeted
from loss import pairwise_similarity, NT_xent
args = linear_parser()
use_cuda = torch.cuda.is_available()
if use_cuda:
ngpus_per_node = torch.cuda.device_count()
if args.local_rank % ngpus_per_node == 0:
print_args(args)
def print_status(string):
if args.local_rank % ngpus_per_node == 0:
print(string)
print_status(torch.cuda.device_count())
print_status('Using CUDA..')
best_acc = 0 # best test accuracy
start_epoch = 0 # start from epoch 0 or last checkpoint epoch
if args.seed != 0:
torch.manual_seed(args.seed)
def main(args):
save_folder = '%s_%s' % (args.dataset, args.affix)
log_folder = os.path.join(args.log_root, save_folder)
model_folder = os.path.join(args.model_root, save_folder)
makedirs(log_folder)
makedirs(model_folder)
setattr(args, 'log_folder', log_folder)
setattr(args, 'model_folder', model_folder)
logger = create_logger(log_folder, args.todo, 'info')
print_args(args, logger)
# model = WideResNet(depth=34, num_classes=10, widen_factor=10, dropRate=0.0)
model = models.resnet50(pretrained=args.pretrain)
num_classes=8
# model.classifier = nn.Linear(model.classifier.in_features, num_classes)
model.fc = nn.Linear(model.fc.in_features, num_classes)
attack = FastGradientSignUntargeted(model,
args.epsilon,
args.alpha,
min_val=0,
max_val=1,
max_iters=args.k,
_type=args.perturbation_type)
if torch.cuda.is_available():
model.cuda()
# model = nn.DataParallel(model).cuda()
trainer = Trainer(args, logger, attack)
if args.todo == 'train':
transform_train = tv.transforms.Compose([
tv.transforms.Resize(256),
tv.transforms.ToTensor(),
tv.transforms.Lambda(lambda x: F.pad(x.unsqueeze(0),
(4*6,4*6,4*6,4*6), mode='constant', value=0).squeeze()),
tv.transforms.ToPILImage(),
tv.transforms.RandomHorizontalFlip(),
tv.transforms.ColorJitter(brightness=0.3, contrast=0.3,
saturation=0.3, hue=0.3),
# tv.transforms.RandomRotation(25),
tv.transforms.RandomAffine(25, translate=(0.2, 0.2),
scale=(0.8,1.2), shear=10),
tv.transforms.RandomCrop(256),
tv.transforms.ToTensor(),
AddGaussianNoise(0.5, args.epsilon)
])
tr_dataset = patd.PatchDataset(path_to_images=args.data_root,
fold='train',
sample=args.subsample,
transform=transform_train)
tr_loader = DataLoader(tr_dataset, batch_size=args.batch_size, shuffle=True, num_workers=24)
# evaluation during training
transform_test = tv.transforms.Compose([
tv.transforms.Resize(256),
# tv.transforms.CenterCrop(224),
tv.transforms.ToTensor(),
# tv.transforms.Normalize(mean, std)
])
te_dataset = patd.PatchDataset(path_to_images=args.data_root,
fold='valid',
transform=transform_test)
te_loader = DataLoader(te_dataset, batch_size=args.batch_size, shuffle=False, num_workers=24)
trainer.train(model, tr_loader, te_loader, args.adv_train)
elif args.todo == 'test':
te_dataset = patd.PatchDataset(path_to_images=args.data_root,
fold='test',
transform=tv.transforms.Compose([
tv.transforms.Resize(256),
tv.transforms.ToTensor(),
]))
te_loader = DataLoader(te_dataset, batch_size=1, shuffle=False, num_workers=1)
checkpoint = torch.load(args.load_checkpoint)
model.load_state_dict(checkpoint)
std_acc, adv_acc = trainer.test(model, te_loader, adv_test=True, use_pseudo_label=False, if_AUC=True)
print("std acc: %.4f, adv_acc: %.4f" % (std_acc * 100, adv_acc * 100))
else:
raise NotImplementedError
def main(args):
save_folder = '%s_%s' % (args.dataset, args.affix)
log_folder = os.path.join(args.log_root, save_folder)
model_folder = os.path.join(args.model_root, save_folder)
makedirs(log_folder)
makedirs(model_folder)
setattr(args, 'log_folder', log_folder)
setattr(args, 'model_folder', model_folder)
logger = create_logger(log_folder, args.todo, 'info')
print_args(args, logger)
# Using a WideResNet model
model = WideResNet(depth=34, num_classes=10, widen_factor=1, dropRate=0.0)
flop, param = get_model_infos(model, (1, 3, 32, 32))
logger.info('Model Info: FLOP = {:.2f} M, Params = {:.2f} MB'.format(
flop, param))
# Configuring the train attack mode
if args.adv_train_mode == 'FGSM':
train_attack = FastGradientSignUntargeted(model,
args.epsilon,
args.alpha,
min_val=0,
max_val=1,
max_iters=args.k,
_type=args.perturbation_type,
logger=logger)
elif args.adv_train_mode == 'CW':
mean = [0]
std = [1]
inputs_box = (min((0 - m) / s for m, s in zip(mean, std)),
max((1 - m) / s for m, s in zip(mean, std)))
train_attack = carlini_wagner_L2.L2Adversary(targeted=False,
confidence=0.0,
search_steps=10,
optimizer_lr=5e-4,
logger=logger)
# Configuring the test attack mode
if args.adv_test_mode == 'FGSM':
test_attack = FastGradientSignUntargeted(model,
args.epsilon,
args.alpha,
min_val=0,
max_val=1,
max_iters=args.k,
_type=args.perturbation_type,
logger=logger)
elif args.adv_test_mode == 'CW':
mean = [0]
std = [1]
inputs_box = (min((0 - m) / s for m, s in zip(mean, std)),
max((1 - m) / s for m, s in zip(mean, std)))
test_attack = carlini_wagner_L2.L2Adversary(targeted=False,
confidence=0.0,
search_steps=10,
optimizer_lr=5e-4,
logger=logger)
if torch.cuda.is_available():
model.cuda()
trainer = Trainer(args, logger, train_attack, test_attack)
if args.todo == 'train':
transform_train = tv.transforms.Compose([
tv.transforms.ToTensor(),
tv.transforms.Lambda(lambda x: F.pad(
x.unsqueeze(0),
(4, 4, 4, 4), mode='constant', value=0).squeeze()),
tv.transforms.ToPILImage(),
tv.transforms.RandomCrop(32),
tv.transforms.RandomHorizontalFlip(),
tv.transforms.ToTensor(),
])
tr_dataset = tv.datasets.CIFAR10(args.data_root,
train=True,
transform=transform_train,
download=True)
tr_loader = DataLoader(tr_dataset,
batch_size=args.batch_size,
shuffle=True,
num_workers=4)
# evaluation during training
te_dataset = tv.datasets.CIFAR10(args.data_root,
train=False,
transform=tv.transforms.ToTensor(),
download=True)
te_loader = DataLoader(te_dataset,
batch_size=args.batch_size,
shuffle=False,
num_workers=4)
trainer.train(model, tr_loader, te_loader, args.adv_train)
elif args.todo == 'test':
pass
else:
raise NotImplementedError
for k in range(args.n_train_data):
file = os.path.join(args.data_folder, 'searchPlay-' + str(k))
dataList = dataProcessor.retrieveData(file)
totalDataList.extend(dataList)
currentModel = 0
trainWorker = NetworkTraining()
trainWorker.train(args.trainepochs, currentModel, totalDataList)
if __name__ == '__main__':
argument.initialize_args()
args = argument.get_args()
logger.initialize_logger(args.log_folder, args.todo, 'info')
logger = logger.get_logger()
timer.init()
argument.print_args(args, logger)
dataProcessor.initSimulator(Board.Board(args.size, args.numberForWin))
args.save_folder = 'test_visual'
if args.todo == 'selfplaytrain':
train(args, logger, dataProcessor)
elif args.todo == 'visualize':
visualize(args, logger, dataProcessor)
elif args.todo == 'experiment':
experiment(args, logger, dataProcessor)
elif args.todo == 'sampledata':
sampledata(args, logger, dataProcessor)
elif args.todo == 'supervisedtrain':
supervisedtrain(args, logger, dataProcessor)
return
elif args.todo == "class_wise":
# Generate class wise error minimizing noise
perturbation.class_wise(train_loader,
num_classes=10,
save=args.delta_path + args.save)
return
else:
raise NotImplementedError
if __name__ == "__main__":
args = argument.parser()
argument.print_args(args)
# Set seed and device
seed = 0
if torch.cuda.is_available() and args.gpu != 0:
device = torch.device("cuda")
torch.cuda.manual_seed(seed)
torch.cuda.manual_seed_all(seed)
else:
device = torch.device("cpu")
torch.manual_seed(seed)
np.random.seed(seed)
random.seed(seed)
torch.backends.cudnn.benchmark = True
