def FER_batch_test():
global PublicTestset
global PublicTestloader
global PrivateTestset
global PrivateTestloader
data_file = './data/Fer2013.h5'
t_length = 28709
v_length = 3589
te_length = 3589
re_length = 96
batchsize = 8
PublicTestset = FER2013(split='PublicTest',
filename=data_file,
train_length=t_length,
validate_length=v_length,
test_length=te_length,
resize_length=re_length,
transform=transform_test)
PublicTestloader = torch.utils.data.DataLoader(PublicTestset,
batch_size=batchsize,
shuffle=False)
PrivateTestset = FER2013(split='PrivateTest',
filename=data_file,
train_length=t_length,
validate_length=v_length,
test_length=te_length,
resize_length=re_length,
transform=transform_test)
PrivateTestloader = torch.utils.data.DataLoader(PrivateTestset,
batch_size=batchsize,
shuffle=False)
detailed_batch_test()
'Angry', 'Disgust', 'Fear', 'Happy', 'Sad', 'Surprise', 'Neutral'
]
# Model
if opt.model == 'VGG19':
net = VGG('VGG19')
elif opt.model == 'Resnet18':
net = ResNet18()
path = os.path.join(opt.dataset + '_' + opt.model)
checkpoint = torch.load(os.path.join(path, opt.split + '_model.t7'))
net.load_state_dict(checkpoint['net'])
net.cuda()
net.eval()
Testset = FER2013(split=opt.split, transform=transform_test)
Testloader = torch.utils.data.DataLoader(Testset,
batch_size=128,
shuffle=False,
num_workers=1)
correct = 0
total = 0
all_target = []
for batch_idx, (inputs, targets) in enumerate(Testloader):
bs, ncrops, c, h, w = np.shape(inputs)
inputs = inputs.view(-1, c, h, w)
inputs, targets = inputs.cuda(), targets.cuda()
inputs, targets = Variable(inputs, volatile=True), Variable(targets)
outputs = net(inputs)
# Data
print('==> Preparing data..')
transform_train = transforms.Compose([
transforms.RandomCrop(44),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
])
transform_test = transforms.Compose([
transforms.TenCrop(cut_size),
transforms.Lambda(lambda crops: torch.stack(
[transforms.ToTensor()(crop) for crop in crops])),
])
trainset = FER2013(split='Training', transform=transform_train)
trainloader = torch.utils.data.DataLoader(trainset,
batch_size=opt.bs,
shuffle=True,
num_workers=1)
PublicTestset = FER2013(split='PublicTest', transform=transform_test)
PublicTestloader = torch.utils.data.DataLoader(PublicTestset,
batch_size=opt.bs,
shuffle=False,
num_workers=1)
PrivateTestset = FER2013(split='PrivateTest', transform=transform_test)
PrivateTestloader = torch.utils.data.DataLoader(PrivateTestset,
batch_size=opt.bs,
shuffle=False,
num_workers=1)
from torch.autograd import Variable
from models import *
from torchvision.models import resnext101_32x8d
use_cuda = torch.cuda.is_available()
cut_size = 44
alpha = 0.5
transform_test = transforms.Compose([
transforms.TenCrop(cut_size),
transforms.Lambda(lambda crops: torch.stack(
[transforms.ToTensor()(crop) for crop in crops])),
])
PrivateTestset = FER2013(split='PrivateTest', transform=transform_test)
PrivateTestloader = torch.utils.data.DataLoader(PrivateTestset,
batch_size=64,
shuffle=False,
num_workers=1)
criterion = nn.CrossEntropyLoss()
def PrivateTest_adv(net):
net.eval()
PrivateTest_loss = 0
correct = 0
total = 0
err0 = 0.005 # for FGSM
for batch_idx, (inputs, targets) in enumerate(PrivateTestloader):
bs, ncrops, c, h, w = np.shape(inputs)
# Using the trained VGG19 model
path = "FER2013_VGG19"
model = VGG("VGG19")
checkpoint = torch.load(os.path.join(path, 'PrivateTest_model.t7'))
model.load_state_dict(checkpoint['net'])
model.to(device)
# Preprocessing the images, sample from training set
cut_size = 44
transform_train = transforms.Compose([
transforms.RandomCrop(44),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
])
trainset = FER2013(split='Training', transform=transform_train)
trainloader = torch.utils.data.DataLoader(trainset,
batch_size=batch_size,
shuffle=True,
num_workers=1)
# sample the 23th batch data
index = 0
for cln_data, true_label in trainloader:
index += 1
if index == 23:
break
bs, c, h, w = np.shape(cln_data)
# print true labels
print(true_label)
])
transform_test = transforms.Compose([
transforms.FiveCrop(cut_size),
transforms.Lambda(lambda crops: torch.stack(
[transforms.ToTensor()(crop) for crop in crops])),
])
# transform_test = transforms.Compose([
# transforms.ToTensor(),
# ])
trainset = FER2013(split='Training',
filename=data_file,
train_length=t_length,
validate_length=v_length,
test_length=te_length,
resize_length=re_length,
transform=transform_train)
trainloader = torch.utils.data.DataLoader(trainset,
batch_size=opt.bs,
shuffle=True)
PublicTestset = FER2013(split='PublicTest',
filename=data_file,
train_length=t_length,
validate_length=v_length,
test_length=te_length,
resize_length=re_length,
transform=transform_test)
PublicTestloader = torch.utils.data.DataLoader(PublicTestset,
batch_size=16,
