# argparse
parser = argparse.ArgumentParser(description='Supervised Multi Layer Perceptron Example')
parser.add_argument('--epoch', '-e', type=int, default=10, help='training epoch (default: 10)')
parser.add_argument('--batch', '-b', type=int, default=300, help='training batchsize (default: 300)')
parser.add_argument('--valbatch', '-v', type=int, default=1000, help='validation batchsize (default: 1000)')
parser.add_argument('--gpu', '-g', type=int, default=-1, help='GPU device #, if you want to use cpu, use -1 (default: -1)')
args = parser.parse_args()
def mnist_preprocess(data):
data['data'] /= 255.
return data
# Logger setup
logger = Logger('MNIST AE',
train_log_mode='TRAIN_LOSS_ONLY',
test_log_mode='TEST_LOSS_ONLY')
# Configure GPU Device
if args.gpu >= 0:
cuda.check_cuda_available()
xp = cuda.cupy if args.gpu >= 0 else np
# loading dataset
dataset = mnist.load()
dim = dataset['train']['data'][0].size
N_train = len(dataset['train']['target'])
N_test = len(dataset['test']['target'])
train_data_dict = {'data':dataset['train']['data'].reshape(N_train, dim).astype(np.float32)}
test_data_dict = {'data':dataset['test']['data'].reshape(N_test, dim).astype(np.float32)}
train_data = DataFeeder(train_data_dict, batchsize=args.batch)
parser.add_argument('--epoch', '-e', type=int, default=3, help='training epoch (default: 3)')
parser.add_argument('--batch', '-b', type=int, default=64, help='training batchsize (default: 64)')
parser.add_argument('--valbatch', '-v', type=int, default=64, help='validation batchsize (default: 64)')
parser.add_argument('--model', '-m', type=str, default='baseline', choices=cnn_models.keys(), help='Model you train (default: baseline)')
parser.add_argument('--output', '-o', type=str, default='mnist_baseline.h5', help='Name of trained model file (default: mnist_baseline.h5)')
parser.add_argument('--gpu', '-g', type=int, default=-1, help='GPU device #, if you want to use cpu, use -1 (default: -1)')
args = parser.parse_args()
def mnist_preprocess(data):
data['data'] /= 255.
data['data'] = data['data'].reshape(1, 28, 28)
return data
# Logger setup
logger = Logger('MNIST CNN')
# Configure GPU Device
if args.gpu >= 0:
cuda.check_cuda_available()
xp = cuda.cupy if args.gpu >= 0 else np
# loading dataset
dataset = mnist.load()
dim = dataset['train']['data'][0].size
N_train = len(dataset['train']['target'])
N_test = len(dataset['test']['target'])
train_data_dict = {'data':dataset['train']['data'].reshape(N_train, dim).astype(np.float32),
'target':dataset['train']['target'].astype(np.int32)}
test_data_dict = {'data':dataset['test']['data'].reshape(N_test, dim).astype(np.float32),
'target':dataset['test']['target'].astype(np.int32)}
parser.add_argument(
'--gpu',
'-g',
type=int,
default=-1,
help='GPU device #, if you want to use cpu, use -1 (default: -1)')
args = parser.parse_args()
def mnist_preprocess(data):
data['data'] /= 255.
return data
# Logger setup
logger = Logger('MNIST MLP')
# Configure GPU Device
if args.gpu >= 0:
cuda.check_cuda_available()
xp = cuda.cupy if args.gpu >= 0 else np
# loading dataset
dataset = mnist.load()
dim = dataset['train']['data'][0].size
N_train = len(dataset['train']['target'])
N_test = len(dataset['test']['target'])
train_data_dict = {
'data': dataset['train']['data'].reshape(N_train, dim).astype(np.float32),
'target': dataset['train']['target'].astype(np.int32)
}
def mnist_preprocess_u(data):
data['data'] /= 255.
return data
# Logger setup
def vat_train_log(res):
log_str = '{0:d}, loss={1:.5f}, lds={2:.5f}'.format(
res['iteration'], res['loss'], res['lds'])
return log_str
logger = Logger('MNIST SIAMESE MLP VAT',
train_log_mode='TRAIN_VAT',
test_log_mode='TEST_LOSS_ONLY')
logger.mode['TRAIN_VAT'] = vat_train_log
# Configure GPU Device
if args.gpu >= 0:
cuda.check_cuda_available()
xp = cuda.cupy if args.gpu >= 0 else np
# loading dataset
dataset = mnist.load()
dim = dataset['train']['data'][0].size
N_train = len(dataset['train']['target'])
N_test = len(dataset['test']['target'])
test_data_dict = {
'data': dataset['test']['data'].reshape(N_test, dim).astype(np.float32),
parser.add_argument('--ubatch', '-u', type=int, default=250, help='unlabeled training batchsize (default: 250)')
parser.add_argument('--valbatch', '-v', type=int, default=1000, help='validation batchsize (default: 1000)')
parser.add_argument('--slabeled', '-s', type=int, default=100, help='size of labeled training samples (default: 100)')
parser.add_argument('--gpu', '-g', type=int, default=-1, help='GPU device #, if you want to use cpu, use -1 (default: -1)')
args = parser.parse_args()
def mnist_preprocess(data):
data['data'] /= 255.
return data
# Logger setup
def vat_train_log(res):
log_str = '{0:d}, loss={1:.5f}, lds={2:.5f}, accuracy={3:.5f}'.format(res['iteration'], res['loss'], res['lds'], res['accuracy'])
return log_str
logger = Logger('MNIST MLP', train_log_mode='TRAIN_VAT')
logger.mode['TRAIN_VAT'] = vat_train_log
# Configure GPU Device
if args.gpu >= 0:
cuda.check_cuda_available()
xp = cuda.cupy if args.gpu >= 0 else np
# loading dataset
dataset = mnist.load()
dim = dataset['train']['data'][0].size
N_train = len(dataset['train']['target'])
N_test = len(dataset['test']['target'])
test_data_dict = {'data':dataset['test']['data'].reshape(N_test, dim).astype(np.float32),
'target':dataset['test']['target'].astype(np.int32)}
unlabeled_data_dict = {'data':dataset['train']['data'].reshape(N_train, dim).astype(np.float32)}
args = parser.parse_args()
def mnist_preprocess(data):
data['data'] /= 255.
return data
# Logger setup
def vat_train_log(res):
log_str = '{0:d}, loss={1:.5f}, lds={2:.5f}, accuracy={3:.5f}'.format(
res['iteration'], res['loss'], res['lds'], res['accuracy'])
return log_str
logger = Logger('MNIST MLP', train_log_mode='TRAIN_VAT')
logger.mode['TRAIN_VAT'] = vat_train_log
# Configure GPU Device
if args.gpu >= 0:
cuda.check_cuda_available()
xp = cuda.cupy if args.gpu >= 0 else np
# loading dataset
dataset = mnist.load()
dim = dataset['train']['data'][0].size
N_train = len(dataset['train']['target'])
N_test = len(dataset['test']['target'])
test_data_dict = {
'data': dataset['test']['data'].reshape(N_test, dim).astype(np.float32),
'target': dataset['test']['target'].astype(np.int32)
# argparse
parser = argparse.ArgumentParser(description='All Convolutional Network Example on CIFAR-10')
parser.add_argument('--epoch', '-e', type=int, default=300, help='training epoch (default: 100)')
parser.add_argument('--batch', '-b', type=int, default=500, help='training batchsize (default: 100)')
parser.add_argument('--valbatch', '-v', type=int, default=1000, help='validation batchsize (default: 100)')
parser.add_argument('--gpu', '-g', type=int, default=-1, help='GPU device #, if you want to use cpu, use -1 (default: -1)')
args = parser.parse_args()
def cifar_preprocess(data):
data['data'] /= 255.
return data
# Logger setup
logger = Logger('CIFAR10 AllConvNet')
# Configure GPU Device
if args.gpu >= 0:
cuda.check_cuda_available()
xp = cuda.cupy if args.gpu >= 0 else np
# loading dataset
dataset = cifar10.load()
dim = dataset['train']['data'][0].size
N_train = len(dataset['train']['target'])
N_test = len(dataset['test']['target'])
train_data_dict = {'data':dataset['train']['data'].astype(np.float32),
'target':dataset['train']['target'].astype(np.int32)}
test_data_dict = {'data':dataset['test']['data'].astype(np.float32),
'-g',
type=int,
default=-1,
help='GPU device #, if you want to use cpu, use -1 (default: -1)')
args = parser.parse_args()
def cifar_preprocess(data):
data['data0'] /= 255.
data['data1'] /= 255.
return data
# Logger setup
logger = Logger('CIFAR SIAMESE',
train_log_mode='TRAIN_LOSS_ONLY',
test_log_mode='TEST_LOSS_ONLY')
# Configure GPU Device
if args.gpu >= 0:
cuda.check_cuda_available()
xp = cuda.cupy if args.gpu >= 0 else np
# loading dataset
dataset = cifar10.load()
dim = dataset['train']['data'][0].size
N_train = len(dataset['train']['target'])
N_test = len(dataset['test']['target'])
train_data_dict = {
'data': dataset['train']['data'].astype(np.float32),