logger("=================FLAGS==================")
for k, v in args.__dict__.items():
logger('{}: {}'.format(k, v))
logger("========================================")
# seed
args.cuda = torch.cuda.is_available()
torch.manual_seed(args.seed)
if args.cuda:
torch.cuda.manual_seed(args.seed)
# data loader and model
assert args.type in ['cifar10', 'cifar100'], args.type
train_loader, test_loader = dataset.get10(batch_size=args.batch_size,
num_workers=1)
model = model.cifar10(args=args, logger=logger)
if args.cuda:
model.cuda()
optimizer = optim.SGD(model.parameters(), lr=1)
decreasing_lr = list(map(int, args.decreasing_lr.split(',')))
logger('decreasing_lr: ' + str(decreasing_lr))
best_acc, old_file = 0, None
t_begin = time.time()
grad_scale = args.grad_scale
try:
# ready to go
for epoch in range(args.epochs):
model.train()
logger('{}: {}'.format(k, v))
logger("========================================")
# seed
args.cuda = torch.cuda.is_available()
torch.manual_seed(args.seed)
if args.cuda:
torch.cuda.manual_seed(args.seed)
model_path = './log/default/batch_size=200/decreasing_lr=200,250/grad_scale=8/seed=117/type=cifar10/wl_activate=8/wl_error=8/wl_grad=8/wl_weight=8/latest.pth'
# data loader and model
assert args.type in ['cifar10', 'cifar100'], args.type
train_loader, test_loader = dataset.get10(batch_size=args.batch_size,
num_workers=1)
modelCF = model.cifar10(args=args, logger=logger, pretrained=model_path)
print(args.cuda)
if args.cuda:
modelCF.cuda()
best_acc, old_file = 0, None
t_begin = time.time()
# ready to go
modelCF.eval()
test_loss = 0
correct = 0
trained_with_quantization = True
# for data, target in test_loader:
for i, (data, target) in enumerate(test_loader):
if i == 0:
hook_handle_list = hook.hardware_evaluation(modelCF, args.wl_weight,
import os
from PIL import ImageOps
from models import CompletionNetwork
from utils import poisson_blend_old
from grad_cam import GradCam
##################################################
# PARAMETER SETTING
##################################################
MODEL = 'cifar10_net.pth'
MASK_COND = 0.7
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
BATCH_SIZE = 32
##################################################
# Load the backdoored model
net = cifar10(128)
net = net.to(device)
net.load_state_dict(torch.load(MODEL))
net.eval()
gcam = GradCam(net, True, device)
print("Loading model successfully\n")
# Data Loader
# --------------
data_transforms = {
'train':
transforms.Compose([
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
]),