# Configure GPU
config_visible_gpu(args.gpu)
use_gpu = args.gpu != 'cpu' and torch.cuda.is_available()
device = torch.device('cuda:0' if use_gpu else 'cpu')
# Data set
if args.dataset.lower() in [
'mnist',
]:
train_loader, test_loader = mnist(batch_size=args.batch_size,
batch_size_test=args.batch_size)
elif args.dataset.lower() in [
'cifar10',
]:
train_loader, test_loader = cifar10(batch_size=args.batch_size,
batch_size_test=args.batch_size)
else:
raise ValueError('Unrecognized dataset: %s' % args.dataset)
assert args.subset in [
'train', 'test'
], 'Subset tag can be only "train" or "test", but %s found' % args.subset
data_loader = train_loader if args.subset in [
'train',
] else test_loader
# Parse IO
assert os.path.exists(
args.model2load
), 'model2load file %s does not exists' % args.model2load
out_dir = os.path.dirname(args.out_file)
if out_dir != '' and not os.path.exists(out_dir):
config_visible_gpu(args.gpu)
use_gpu = args.gpu != 'cpu' and torch.cuda.is_available()
device = torch.device('cuda:0' if use_gpu else 'cpu')
# Data set
if args.dataset.lower() in [
'mnist',
]:
train_loader, test_loader = mnist(batch_size=args.batch_size,
batch_size_test=args.batch_size,
normalization=args.normalization)
elif args.dataset.lower() in [
'cifar10',
]:
train_loader, test_loader = cifar10(batch_size=args.batch_size,
batch_size_test=args.batch_size,
normalization=args.normalization)
else:
raise ValueError('Unrecognized dataset: %s' % args.dataset)
# Parse IO
if not os.path.exists(args.out_folder):
os.makedirs(args.out_folder)
# Parse model
model = MLP(in_dim=args.in_dim,
hidden_dims=args.hidden_dims,
out_dim=args.out_dim,
nonlinearity=args.nonlinearity)
criterion = nn.CrossEntropyLoss()
model = model.cuda(device) if use_gpu else model
type=str,
default=None,
help='The folder to be scanned')
args = parser.parse_args()
if args.dataset.lower() in [
'mnist',
]:
train_loader, test_loader, classes = mnist(batch_size=args.batch_size,
shuffle=False,
data_augmentation=False)
elif args.dataset.lower() in [
'cifar10',
]:
train_loader, test_loader, classes = cifar10(
batch_size=args.batch_size, shuffle=False, data_augmentation=False)
else:
raise ValueError('Unrecognized dataset: %s' % args.dataset)
train_ori_data = []
test_ori_data = []
for idx, (data_batch, label_batch) in enumerate(train_loader, 0):
data_batch = data_batch.reshape(data_batch.size(0), -1)
train_ori_data.append(data_batch.data.cpu().numpy())
for idx, (data_batch, label_batch) in enumerate(test_loader, 0):
data_batch = data_batch.reshape(data_batch.size(0), -1)
test_ori_data.append(data_batch.data.cpu().numpy())
train_ori_data = np.concatenate(train_ori_data, axis=0)
test_ori_data = np.concatenate(test_ori_data, axis=0)
args = parser.parse_args()
# Configure GPU
config_visible_gpu(args.gpu)
use_gpu = args.gpu != 'cpu' and torch.cuda.is_available()
device = torch.device('cuda:0' if use_gpu else 'cpu')
# Data set
if args.dataset.lower() in [
'mnist',
]:
train_loader, test_loader = mnist(batch_size=1, batch_size_test=1)
elif args.dataset.lower() in [
'cifar10',
]:
train_loader, test_loader = cifar10(batch_size=1, batch_size_test=1)
else:
raise ValueError('Unrecognized dataset: %s' % args.dataset)
assert args.subset in [
'train', 'test'
], 'Subset tag can be only "train" or "test", but %s found' % args.subset
data_loader = train_loader if args.subset in [
'train',
] else test_loader
# Parse IO
assert os.path.exists(
args.model2load
), 'model2load file %s does not exists' % args.model2load
out_dir = os.path.dirname(args.out_file)
if out_dir != '' and not os.path.exists(out_dir):