557 lines
28 KiB
Python
557 lines
28 KiB
Python
|
import os
|
||
|
|
import sys
|
||
|
|
sys.path.insert(0, '../')
|
||
|
|
import numpy as np
|
||
|
|
import torch
|
||
|
|
import logging
|
||
|
|
import argparse
|
||
|
|
import torch.nn as nn
|
||
|
|
from PIL import Image
|
||
|
|
import torch.utils
|
||
|
|
import torchvision.datasets as dset
|
||
|
|
import torch.backends.cudnn as cudnn
|
||
|
|
import torchvision.transforms as transforms
|
||
|
|
from torch.autograd import Variable
|
||
|
|
from retrain import utils
|
||
|
|
from retrain.model import NetworkCIFAR as Network
|
||
|
|
from retrain import genotypes
|
||
|
|
from AAA.eval_attack.eval_manual import Normalize
|
||
|
|
from noise.jacobian import JacobianReg
|
||
|
|
from noise.CAA_noise.attacker_small import CAA
|
||
|
|
os.environ['MPLCONFIGDIR'] = os.getcwd() + "/configs/"
|
||
|
|
import matplotlib.pyplot as plt
|
||
|
|
import seaborn as sns
|
||
|
|
import time
|
||
|
|
import glob
|
||
|
|
|
||
|
|
parser = argparse.ArgumentParser("cifar")
|
||
|
|
parser.add_argument('--data', type=str, default='/mnt/jfs/sunjialiang/data', help='location of the data corpus')
|
||
|
|
parser.add_argument('--set', type=str, default='cifar10', help='location of the data corpus')
|
||
|
|
parser.add_argument('--batch_size', type=int, default=32, help='batch size')
|
||
|
|
parser.add_argument('--learning_rate', type=float, default=0.025, help='init learning rate')
|
||
|
|
parser.add_argument('--momentum', type=float, default=0.9, help='momentum')
|
||
|
|
parser.add_argument('--weight_decay', type=float, default=3e-4, help='weight decay')
|
||
|
|
parser.add_argument('--report_freq', type=float, default=200, help='report frequency')
|
||
|
|
# parser.add_argument('--gpu', type=int, default=3, help='gpu device id')
|
||
|
|
parser.add_argument('--epochs', type=int, default=50, help='num of training epochs')
|
||
|
|
parser.add_argument('--init_channels', type=int, default=36, help='num of init channels')
|
||
|
|
parser.add_argument('--layers', type=int, default=20, help='total number of layers')
|
||
|
|
parser.add_argument('--model_path', type=str, default='saved_models', help='path to save the model')
|
||
|
|
parser.add_argument('--auxiliary', action='store_true', default=False, help='use auxiliary tower')
|
||
|
|
parser.add_argument('--auxiliary_weight', type=float, default=0.4, help='weight for auxiliary loss')
|
||
|
|
parser.add_argument('--cutout', action='store_true', default=False, help='use cutout')
|
||
|
|
parser.add_argument('--max_epsilon', type=float, default=0.031, help='the attack sequence length')
|
||
|
|
parser.add_argument('--norm', default='linf', help='linf | l2 | unrestricted')
|
||
|
|
parser.add_argument('--cutout_length', type=int, default=16, help='cutout length')
|
||
|
|
parser.add_argument('--drop_path_prob', type=float, default=0.3, help='drop path probability')
|
||
|
|
parser.add_argument('--save', type=str, default='EXP', help='experiment name')
|
||
|
|
parser.add_argument('--seed', type=int, default=0, help='random seed')
|
||
|
|
parser.add_argument('--arch', type=str, default='DARTS_total', help='which architecture to use')
|
||
|
|
parser.add_argument('--grad_clip', type=float, default=5, help='gradient clipping')
|
||
|
|
args = parser.parse_args()
|
||
|
|
|
||
|
|
class CorruptDataset(torch.utils.data.Dataset):
|
||
|
|
# '''
|
||
|
|
# In CIFAR-10-C, the first 10,000 images in each .npy are the test set images corrupted at severity 1,
|
||
|
|
# and the last 10,000 images are the test set images corrupted at severity five. labels.npy is the label file for all other image files.
|
||
|
|
# '''
|
||
|
|
def __init__(self, transform):
|
||
|
|
images = np.load('/mnt/jfs/sunjialiang/AAAD/noise/CIFAR-10-C/snow.npy')
|
||
|
|
labels = np.load('/mnt/jfs/sunjialiang/AAAD/noise/CIFAR-10-C/labels.npy')
|
||
|
|
assert labels.min() >= 0
|
||
|
|
assert images.dtype == np.uint8
|
||
|
|
assert images.shape[0] <= 50000
|
||
|
|
assert images.shape[1:] == (32, 32, 3)
|
||
|
|
self.images = [Image.fromarray(x) for x in images]
|
||
|
|
# self.labels = labels / labels.sum(axis=1, keepdims=True) # normalize
|
||
|
|
self.labels = labels.astype(np.float32)
|
||
|
|
self.transform = transform
|
||
|
|
|
||
|
|
def __getitem__(self, index):
|
||
|
|
image, label = self.images[index], self.labels[index]
|
||
|
|
image = self.transform(image)
|
||
|
|
return image, label
|
||
|
|
|
||
|
|
def __len__(self):
|
||
|
|
return len(self.labels)
|
||
|
|
|
||
|
|
class SystemDataset(torch.utils.data.Dataset):
|
||
|
|
def __init__(self, transform):
|
||
|
|
images = np.load('/mnt/jfs/sunjialiang/AAAD/noise/CIFAR-10-S/opencv-pil_bilinear.npy')
|
||
|
|
labels = np.load('/mnt/jfs/sunjialiang/AAAD/noise/CIFAR-10-S/label_clean.npy')
|
||
|
|
assert labels.min() >= 0
|
||
|
|
assert images.dtype == np.uint8
|
||
|
|
assert images.shape[0] <= 60000
|
||
|
|
assert images.shape[1:] == (32, 32, 3)
|
||
|
|
self.images = [Image.fromarray(x) for x in images]
|
||
|
|
# self.labels = labels / labels.sum(axis=1, keepdims=True) # normalize
|
||
|
|
self.labels = labels.astype(np.float32)
|
||
|
|
self.transform = transform
|
||
|
|
|
||
|
|
def __getitem__(self, index):
|
||
|
|
image, label = self.images[index], self.labels[index]
|
||
|
|
image = self.transform(image)
|
||
|
|
return image, label
|
||
|
|
|
||
|
|
def __len__(self):
|
||
|
|
return len(self.labels)
|
||
|
|
|
||
|
|
def ra_System(model,valid_queue):
|
||
|
|
transform_valid = transforms.Compose([
|
||
|
|
transforms.RandomCrop(32, padding=4),
|
||
|
|
transforms.RandomHorizontalFlip(),
|
||
|
|
transforms.ToTensor(),
|
||
|
|
transforms.Normalize((0.49139968, 0.48215827, 0.44653124), (0.24703233, 0.24348505, 0.26158768)),
|
||
|
|
])
|
||
|
|
data = SystemDataset(transform=transform_valid)
|
||
|
|
num_data = len(data)
|
||
|
|
indices = list(range(num_data))
|
||
|
|
split = int(np.floor(50001))
|
||
|
|
valid_queue = torch.utils.data.DataLoader(
|
||
|
|
data, batch_size=64, sampler=torch.utils.data.sampler.SubsetRandomSampler(indices[split:num_data]),shuffle=False, pin_memory=True, num_workers=2)
|
||
|
|
top1_clean = utils.AvgrageMeter()
|
||
|
|
model.eval()
|
||
|
|
for step, (input, target) in enumerate(valid_queue):
|
||
|
|
target = target.argmax(dim=1)
|
||
|
|
input = Variable(input, volatile=True).cuda()
|
||
|
|
target = Variable(target, volatile=True).cuda()
|
||
|
|
logits_clean = model(input)
|
||
|
|
# visualization
|
||
|
|
# images_v = input.squeeze(0)
|
||
|
|
# print(input.shape)
|
||
|
|
# plt.axis('off')
|
||
|
|
# plt.imshow(images_v.permute(1, 2, 0).cpu().numpy())
|
||
|
|
# plt.show()
|
||
|
|
prec1_clean, _ = utils.accuracy(logits_clean, target, topk=(1, 5))
|
||
|
|
n = input.size(0)
|
||
|
|
top1_clean.update(prec1_clean.item(), n)
|
||
|
|
# if step % args.report_freq == 0:
|
||
|
|
# logging.info('valid %03d %f', step, top1_clean.avg)
|
||
|
|
return top1_clean.avg
|
||
|
|
|
||
|
|
def ra_corruption(model,valid_queue):
|
||
|
|
transform_valid = transforms.Compose([
|
||
|
|
transforms.RandomCrop(32, padding=4),
|
||
|
|
transforms.RandomHorizontalFlip(),
|
||
|
|
transforms.ToTensor(),
|
||
|
|
transforms.Normalize((0.49139968, 0.48215827, 0.44653124), (0.24703233, 0.24348505, 0.26158768)),
|
||
|
|
])
|
||
|
|
data = CorruptDataset(transform=transform_valid)
|
||
|
|
num_data = len(data)
|
||
|
|
indices = list(range(num_data))
|
||
|
|
split = int(np.floor(10000))
|
||
|
|
valid_queue = torch.utils.data.DataLoader(
|
||
|
|
data, batch_size=64, sampler=torch.utils.data.sampler.SubsetRandomSampler(indices[split:split+10000]),shuffle=False, pin_memory=True, num_workers=2)
|
||
|
|
top1_clean = utils.AvgrageMeter()
|
||
|
|
model.eval()
|
||
|
|
for step, (input, target) in enumerate(valid_queue):
|
||
|
|
input = Variable(input, volatile=True).cuda()
|
||
|
|
target = Variable(target, volatile=True).cuda()
|
||
|
|
logits_clean = model(input)
|
||
|
|
# visualization
|
||
|
|
# images_v = input[0,:,:,:].squeeze(0)
|
||
|
|
# print(input.shape)
|
||
|
|
# plt.axis('off')
|
||
|
|
# plt.imshow(images_v.permute(1, 2, 0).cpu().numpy())
|
||
|
|
# plt.show()
|
||
|
|
prec1_clean, _ = utils.accuracy(logits_clean, target, topk=(1, 5))
|
||
|
|
n = input.size(0)
|
||
|
|
top1_clean.update(prec1_clean.item(), n)
|
||
|
|
# if step % args.report_freq == 0:
|
||
|
|
# logging.info('valid %03d %f', step, top1_clean.avg)
|
||
|
|
return top1_clean.avg
|
||
|
|
|
||
|
|
def Jacobian(model,valid_queue):
|
||
|
|
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
||
|
|
reg = JacobianReg()
|
||
|
|
R_jacobian = 0
|
||
|
|
count = 0
|
||
|
|
for idx, (images, labels) in enumerate(valid_queue):
|
||
|
|
data, target = images.to(device), labels.to(device)
|
||
|
|
batch_size = len(images)
|
||
|
|
data.requires_grad = True # this is essential!
|
||
|
|
output = model(data) # forward pass
|
||
|
|
R = reg(data, output) # Jacobian regularization
|
||
|
|
R = R.detach()
|
||
|
|
R_jacobian = R_jacobian + R
|
||
|
|
count = count + 1
|
||
|
|
number = batch_size * count
|
||
|
|
del output, R, images, labels
|
||
|
|
torch.cuda.empty_cache()
|
||
|
|
Jacobian_value = R_jacobian/number
|
||
|
|
# print(Jacobian_value)
|
||
|
|
return Jacobian_value
|
||
|
|
|
||
|
|
def RA_infer(model,valid_queue, args, subpolicy):
|
||
|
|
top1_adv = utils.AvgrageMeter()
|
||
|
|
model.eval()
|
||
|
|
for step, (input, target) in enumerate(valid_queue):
|
||
|
|
input = Variable(input, volatile=True).cuda()
|
||
|
|
target = Variable(target, volatile=True).cuda()
|
||
|
|
logits_clean = model(input)
|
||
|
|
images = CAA(model, args, subpolicy, input, target)
|
||
|
|
logits = model(images)
|
||
|
|
# visualization
|
||
|
|
# images_v = images.squeeze(0)
|
||
|
|
# print(images.shape)
|
||
|
|
# plt.axis('off')
|
||
|
|
# plt.imshow(images_v.permute(1, 2, 0).cpu().numpy())
|
||
|
|
# plt.show()
|
||
|
|
prec1_adv, _ = utils.accuracy(logits, target, topk=(1, 5))
|
||
|
|
n = input.size(0)
|
||
|
|
# print('Robust Accuracy:', top1_adv.avg)
|
||
|
|
top1_adv.update(prec1_adv.item(), n)
|
||
|
|
# if step % args.report_freq == 0:
|
||
|
|
# logging.info('valid %03d %f', step, top1_adv.avg)
|
||
|
|
return top1_adv.avg
|
||
|
|
|
||
|
|
def Clean_infer(model,valid_queue, args):
|
||
|
|
top1_clean = utils.AvgrageMeter()
|
||
|
|
model.eval()
|
||
|
|
for step, (input, target) in enumerate(valid_queue):
|
||
|
|
input = Variable(input, volatile=True).cuda()
|
||
|
|
target = Variable(target, volatile=True).cuda()
|
||
|
|
logits_clean = model(input)
|
||
|
|
# visualization
|
||
|
|
# images_v = input.squeeze(0)
|
||
|
|
# print(input.shape)
|
||
|
|
# plt.axis('off')
|
||
|
|
# plt.imshow(images_v.permute(1, 2, 0).cpu().numpy())
|
||
|
|
# plt.show()
|
||
|
|
prec1_clean, _ = utils.accuracy(logits_clean, target, topk=(1, 5))
|
||
|
|
n = input.size(0)
|
||
|
|
top1_clean.update(prec1_clean.item(), n)
|
||
|
|
# if step % args.report_freq == 0:
|
||
|
|
# logging.info('valid %03d %f', step, top1_clean.avg)
|
||
|
|
return top1_clean.avg
|
||
|
|
|
||
|
|
if __name__ == '__main__':
|
||
|
|
if not torch.cuda.is_available():
|
||
|
|
logging.info('no gpu device available')
|
||
|
|
sys.exit(1)
|
||
|
|
args.save = 'experiments/eval-{}'.format(time.strftime("%Y%m%d-%H%M%S"))
|
||
|
|
utils.create_exp_dir(args.save, scripts_to_save=glob.glob('*.py'))
|
||
|
|
log_format = '%(asctime)s %(message)s'
|
||
|
|
logging.basicConfig(stream=sys.stdout, level=logging.INFO,
|
||
|
|
format=log_format, datefmt='%m/%d %I:%M:%S %p')
|
||
|
|
fh = logging.FileHandler(os.path.join(args.save, 'log.txt'))
|
||
|
|
fh.setFormatter(logging.Formatter(log_format))
|
||
|
|
logging.getLogger().addHandler(fh)
|
||
|
|
|
||
|
|
np.random.seed(args.seed)
|
||
|
|
# torch.cuda.set_device(args.gpu)
|
||
|
|
cudnn.benchmark = True
|
||
|
|
torch.manual_seed(args.seed)
|
||
|
|
cudnn.enabled = True
|
||
|
|
torch.cuda.manual_seed(args.seed)
|
||
|
|
# logging.info('gpu device = %d' % args.gpu)
|
||
|
|
CIFAR_CLASSES = 10
|
||
|
|
criterion = nn.CrossEntropyLoss()
|
||
|
|
criterion = criterion.cuda()
|
||
|
|
if args.set == 'cifar100':
|
||
|
|
train_data = dset.CIFAR100(root=args.data, train=True, download=True, transform=transforms.ToTensor())
|
||
|
|
valid_data = dset.CIFAR100(root=args.data, train=False, download=True, transform=transforms.ToTensor())
|
||
|
|
else:
|
||
|
|
train_data = dset.CIFAR10(root=args.data, train=True, download=True, transform=transforms.ToTensor())
|
||
|
|
valid_data = dset.CIFAR10(root=args.data, train=False, download=True, transform=transforms.ToTensor())
|
||
|
|
train_queue = torch.utils.data.DataLoader(
|
||
|
|
train_data, batch_size=args.batch_size, shuffle=True, pin_memory=True, num_workers=2)
|
||
|
|
valid_queue = torch.utils.data.DataLoader(
|
||
|
|
valid_data, batch_size=16, shuffle=False, pin_memory=True, num_workers=2)
|
||
|
|
|
||
|
|
# model_list = ['PCDART_FGSM_1_100', 'DARTS_V2', 'DARTS_FGSM', 'PCDARTS_natural', 'DARTS_natural', 'NASP_Jacobian', 'DARTS_Arch0', 'dual_adv_fgsm', 'NASP_natural']
|
||
|
|
num_list = []
|
||
|
|
num_list1 = []
|
||
|
|
num_list2 = []
|
||
|
|
for eps in [1/255, 2/255, 3/255]:
|
||
|
|
# model_list = ['DARTS_FGSM','DARTS_PGD','DARTS_natural','DARTS_System','DARTS_Jacobian','DARTS_Hessian']
|
||
|
|
# model_list = ['PCDARTS_FGSM', 'PCDARTS_PGD', 'PCDARTS_natural', 'PCDARTS_System', 'PCDARTS_Jacobian',
|
||
|
|
# 'PCDARTS_Hessian']
|
||
|
|
# model_list = ['SmoothDARTS_FGSM', 'SmoothDARTS_PGD', 'SmoothDARTS_natural', 'SmoothDARTS_System', 'SmoothDARTS_Jacobian',
|
||
|
|
# 'SmoothDARTS_Hessian']
|
||
|
|
# model_list = ['NASP_FGSM', 'NASP_PGD', 'NASP_natural', 'NASP_System', 'NASP_Jacobian',
|
||
|
|
# 'NASP_Hessian']
|
||
|
|
# model_list = ['FairDARTS_FGSM', 'FairDARTS_PGD', 'FairDARTS_natural', 'FairDARTS_System', 'FairDARTS_Jacobian',
|
||
|
|
# 'FairDARTS_Hessian']
|
||
|
|
# model_list = ['Random_search_FGSM', 'Random_search_PGD', 'Random_search_natural', 'Random_search_System', 'Random_search_Jacobian',
|
||
|
|
# 'Random_search_Hessian']
|
||
|
|
# model_list = ['Random_search_weight_sharing_FGSM', 'Random_search_weight_sharing_PGD',
|
||
|
|
# 'Random_search_weight_sharing_natural', 'Random_search_weight_sharing_System',
|
||
|
|
# 'Random_search_weight_sharing_Jacobian',
|
||
|
|
# 'Random_search_weight_sharing_Hessian']
|
||
|
|
model_list = ['DE_FGSM', 'DE_PGD', 'DE_natural', 'DE_System', 'DE_Jacobian',
|
||
|
|
'DE_Hessian']
|
||
|
|
for i in model_list:
|
||
|
|
args.arch = i
|
||
|
|
print(args.arch)
|
||
|
|
genotype = eval("genotypes.%s" % args.arch)
|
||
|
|
model = Network(args.init_channels, CIFAR_CLASSES, args.layers, args.auxiliary, genotype)
|
||
|
|
model = model.cuda()
|
||
|
|
model.drop_path_prob = args.drop_path_prob * 49 / args.epochs
|
||
|
|
# logging.info("param size = %fMB", utils.count_parameters_in_MB(model))
|
||
|
|
if args.arch == 'DARTS_natural':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/DARTS_natural-20221205-222941',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'DARTS_FGSM':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/DARTS_FGSM-20221205-143208',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'DARTS_Jacobian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/DARTS_Jacobian-20221205-223038',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'DARTS_System':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/DARTS_System-20221205-222249',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'DARTS_PGD':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/DARTS_PGD-20221206-091501',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'DARTS_Hessian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/DARTS_Hessian-20221205-182713',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'PCDARTS_natural':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/PCDARTS_natural-20221206-091652',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'PCDARTS_FGSM':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/PCDARTS_FGSM-20221206-011540',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'PCDARTS_Jacobian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/PCDARTS_Jacobian-20221206-140532',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'PCDARTS_System':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/PCDARTS_System-20221206-011613',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'PCDARTS_PGD':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/PCDARTS_PGD-20221206-150913',
|
||
|
|
'weights.pt')
|
||
|
|
|
||
|
|
elif args.arch == 'PCDARTS_Hessian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/PCDARTS_Hessian-20221206-032415',
|
||
|
|
'weights.pt')
|
||
|
|
|
||
|
|
elif args.arch == 'NASP_natural':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/NASP_natural-20221207-033420',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'NASP_FGSM':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/NASP_FGSM-20221206-122250',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'NASP_Jacobian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/NASP_Jacobian-20221206-214806',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'NASP_System':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/NASP_System-20221206-083537',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'NASP_PGD':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/NASP_PGD-20221206-202438',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'NASP_Hessian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/NASP_Hessian-20221207-020710',
|
||
|
|
'weights.pt')
|
||
|
|
|
||
|
|
elif args.arch == 'FairDARTS_natural':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/FairDARTS_natural-20221207-025850',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'FairDARTS_FGSM':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/FairDARTS_FGSM-20221207-105753',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'FairDARTS_Jacobian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/FairDARTS_Jacobian-20221207-143650',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'FairDARTS_System':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/FairDARTS_System-20221207-120855',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'FairDARTS_PGD':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/FairDARTS_PGD-20221207-120755',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'FairDARTS_Hessian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/FairDARTS_Hessian-20221207-060103',
|
||
|
|
'weights.pt')
|
||
|
|
|
||
|
|
elif args.arch == 'Random_search_natural':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/Random_search_natural-20221212-015118',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'Random_search_FGSM':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/Random_search_FGSM-20221208-012311',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'Random_search_Jacobian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/Random_search_Jacobian-20221212-162206',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'Random_search_System':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/Random_search_System-20221213-234338',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'Random_search_PGD':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/Random_search_PGD-20221207-211953',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'Random_search_Hessian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/Random_search_Hessian-20221213-234431',
|
||
|
|
'weights.pt')
|
||
|
|
|
||
|
|
elif args.arch == 'Random_search_weight_sharing_natural':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/Random_search_weight_sharing_natural-20221214-165024',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'Random_search_weight_sharing_FGSM':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/Random_search_weight_sharing_FGSM-20221214-234838',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'Random_search_weight_sharing_Jacobian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/Random_search_weight_sharing_Jacobian-20221218-150842',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'Random_search_weight_sharing_System':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/Random_search_weight_sharing_System-20221214-164952',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'Random_search_weight_sharing_PGD':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/Random_search_weight_sharing_PGD-20221218-150821',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'Random_search_weight_sharing_Hessian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/Random_search_weight_sharing_Hessian-20221219-155616',
|
||
|
|
'weights.pt')
|
||
|
|
|
||
|
|
elif args.arch == 'SmoothDARTS_natural':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/SmoothDARTS_natural-20221207-154051',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'SmoothDARTS_FGSM':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/SmoothDARTS_FGSM-20221207-170429',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'SmoothDARTS_Jacobian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/SmoothDARTS_Jacobian-20221207-211804',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'SmoothDARTS_System':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/SmoothDARTS_System-20221207-211900',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'SmoothDARTS_PGD':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/SmoothDARTS_PGD-20221207-171400',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'SmoothDARTS_Hessian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/SmoothDARTS_Hessian-20221207-131829',
|
||
|
|
'weights.pt')
|
||
|
|
|
||
|
|
elif args.arch == 'DE_natural':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/DE_natural-20221214-234943',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'DE_FGSM':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/DE_FGSM-20221219-150040',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'DE_Jacobian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/DE_Jacobian-20221223-231232',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'DE_System':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/DE_System-20221218-071004',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'DE_PGD':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/DE_PGD-20221223-231438',
|
||
|
|
'weights.pt')
|
||
|
|
elif args.arch == 'DE_Hessian':
|
||
|
|
save_model_path = os.path.join(
|
||
|
|
'/mnt/jfs/sunjialiang/AAAD/retrain/standard_train/trained_model/DE_Hessian-20221230-214437',
|
||
|
|
'weights.pt')
|
||
|
|
|
||
|
|
model.cuda()
|
||
|
|
model.load_state_dict(torch.load(save_model_path))
|
||
|
|
model = nn.Sequential(Normalize([0.49139968, 0.48215827, 0.44653124], [0.24703233, 0.24348505, 0.26158768]),
|
||
|
|
model)
|
||
|
|
model = model.eval()
|
||
|
|
|
||
|
|
subpolicy_linf = [{'attacker': 'GradientSignAttack', 'magnitude': eps, 'step': 1}]
|
||
|
|
subpolicy = subpolicy_linf
|
||
|
|
Ra_adv_FGSM = RA_infer(model, valid_queue, args, subpolicy)
|
||
|
|
logging.info('Attacker %s', subpolicy)
|
||
|
|
logging.info('Robust Accuracy under FGSM = %f', Ra_adv_FGSM)
|
||
|
|
|
||
|
|
if eps == 1/255:
|
||
|
|
num_list.append(Ra_adv_FGSM)
|
||
|
|
elif eps == 2/255:
|
||
|
|
num_list1.append(Ra_adv_FGSM)
|
||
|
|
elif eps == 3/255:
|
||
|
|
num_list2.append(Ra_adv_FGSM)
|
||
|
|
|
||
|
|
# DARTS
|
||
|
|
# num_list = [57.99,59.71,54.75,58.46,60.68,58.25]
|
||
|
|
# num_list1 = [24.79,21.13,21.88,22.48,20.72,22.13]
|
||
|
|
# num_list2 = [17.57,13.36,15.23,15.04,9.79,14.4]
|
||
|
|
|
||
|
|
#Random_search_weight_sharing
|
||
|
|
# num_list = [55.46,63.26,66.61,60.71,59.96,57.84]
|
||
|
|
# num_list1 = [19.49,20.91,20.67,20.61,22.00,20.26]
|
||
|
|
# num_list2 = [13.30,11.71,10.51,12.50,13.99,14.12]
|
||
|
|
|
||
|
|
# PCDARTS
|
||
|
|
# num_list = [58.12,56.24,58.70,57.15,57.17,61.95]
|
||
|
|
# num_list1 = [23.78,21.02,22.22,23.02,20.90,24.68]
|
||
|
|
# num_list2 = [16.08,15.32,14.71,17.55,11.32,16.71]
|
||
|
|
|
||
|
|
# SmoothDARTS
|
||
|
|
# num_list = [62.86,59.19,61.34,56.76,57.63,56.26]
|
||
|
|
# num_list1 = [19.42,17.45,22.65,20.94,23.56,16.62]
|
||
|
|
# num_list2 = [11.30,10.40,13.16,13.75,17.88,9.90]
|
||
|
|
logging.info(num_list)
|
||
|
|
logging.info(num_list1)
|
||
|
|
logging.info(num_list2)
|
||
|
|
sns.set(color_codes=True)
|
||
|
|
name_list = ['FGSM', 'PGD', 'Natural', 'System', 'Jacobian', 'Hessian']
|
||
|
|
x = list(range(len(num_list)))
|
||
|
|
plt.figure(figsize=(12, 12))
|
||
|
|
# plt.figure(figsize=(6,6))
|
||
|
|
total_width, n = 0.7, 3
|
||
|
|
width = total_width / n
|
||
|
|
plt.bar(x, num_list, width=width, label="$\epsilon=1/255$", fc="#0087cb")
|
||
|
|
for i in range(len(x)):
|
||
|
|
x[i] = x[i] + width
|
||
|
|
plt.bar(x, num_list1, width=width, label="$\epsilon=2/255$", tick_label=name_list, fc="#ffa200")
|
||
|
|
for i in range(len(x)):
|
||
|
|
x[i] = x[i] + width
|
||
|
|
plt.bar(x, num_list2, width=width, label="$\epsilon=3/255$", fc="#9966ff")
|
||
|
|
plt.xlabel("Evaluation in search", fontsize=20)
|
||
|
|
plt.ylabel("RA(%)", fontsize=20)
|
||
|
|
plt.title("Evaluation after search: FGSM", fontsize=20)
|
||
|
|
plt.xticks(fontsize=20)
|
||
|
|
plt.yticks(fontsize=20)
|
||
|
|
plt.legend(fontsize=14, loc='upper right')
|
||
|
|
plt.savefig('DE_FGSM.png', bbox_inches='tight', pad_inches=0.1)
|
||
|
|
|
||
|
|
|
||
|
|
|