def prepare_db(opt):
print("Use %s dataset" % (opt.dataset))
if opt.dataset == 'mnist':
train_dataset = torchvision.datasets.MNIST('./data/MNIST', train=True, download=True,
transform=torchvision.transforms.Compose([
torchvision.transforms.ToTensor(),
torchvision.transforms.Normalize((0.1307,), (0.3081,))
]))
eval_dataset = torchvision.datasets.MNIST('./data/MNIST', train=False, download=True,
transform=torchvision.transforms.Compose([
torchvision.transforms.ToTensor(),
torchvision.transforms.Normalize((0.1307,), (0.3081,))
]))
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'adult':
train_dataset = dataset.UCIAdult('./data/uci_adult', train=True)
eval_dataset = dataset.UCIAdult('./data/uci_adult', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'letter':
train_dataset = dataset.UCILetter('./data/uci_letter', train=True)
eval_dataset = dataset.UCILetter('./data/uci_letter', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'yeast':
train_dataset = dataset.UCIYeast('./data/uci_yeast', train=True)
eval_dataset = dataset.UCIYeast('./data/uci_yeast', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
else:
raise NotImplementedError
def prepare_db(opt):
print("Use %s dataset" % (opt.dataset))
if opt.dataset == 'mnist':
train_dataset = torchvision.datasets.MNIST(
'./data/mnist',
train=True,
download=True,
transform=torchvision.transforms.Compose([
torchvision.transforms.ToTensor(),
torchvision.transforms.Normalize((0.1307, ), (0.3081, ))
]))
eval_dataset = torchvision.datasets.MNIST(
'./data/mnist',
train=False,
download=True,
transform=torchvision.transforms.Compose([
torchvision.transforms.ToTensor(),
torchvision.transforms.Normalize((0.1307, ), (0.3081, ))
]))
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'adult':
train_dataset = dataset.UCIAdult('./data/uci_adult', train=True)
eval_dataset = dataset.UCIAdult('./data/uci_adult', train=False)
#print(train_dataset)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'letter':
train_dataset = dataset.UCILetter('./data/uci_letter', train=True)
eval_dataset = dataset.UCILetter('./data/uci_letter', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'yeast':
train_dataset = dataset.UCIYeast('./data/uci_yeast', train=True)
eval_dataset = dataset.UCIYeast('./data/uci_yeast', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'credit':
train_dataset = dataset.UCICredit(
'/home/ubuntu/amberhsia/Neural-Decision-Forests-master/data/q4credit.csv',
train=True)
eval_dataset = dataset.UCICredit(
'/home/ubuntu/amberhsia/Neural-Decision-Forests-master/data/q4credit.csv',
train=False)
return {'train': train_dataset, 'eval': eval_dataset}
else:
raise NotImplementedError
def prepare_db(opt):
print("Use %s dataset" % (opt.dataset))
if opt.dataset == 'mnist':
train_dataset = torchvision.datasets.MNIST('./data/mnist', train=True, download=True,
transform=torchvision.transforms.Compose([
torchvision.transforms.ToTensor(),
torchvision.transforms.Normalize((0.1307,), (0.3081,))
]))
eval_dataset = torchvision.datasets.MNIST('./data/mnist', train=False, download=True,
transform=torchvision.transforms.Compose([
torchvision.transforms.ToTensor(),
torchvision.transforms.Normalize((0.1307,), (0.3081,))
]))
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'adult':
train_dataset = dataset.UCIAdult('./data/uci_adult', train=True)
eval_dataset = dataset.UCIAdult('./data/uci_adult', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'letter':
train_dataset = dataset.UCILetter('./data/uci_letter', train=True)
eval_dataset = dataset.UCILetter('./data/uci_letter', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'yeast':
train_dataset = dataset.UCIYeast('./data/uci_yeast', train=True)
eval_dataset = dataset.UCIYeast('./data/uci_yeast', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'gisette':
train_dataset = dataset.UCIGisette('./data/uci_gisette', train=True)
eval_dataset = dataset.UCIGisette('./data/uci_gisette', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'arrhythmia':
train_dataset = dataset.UCIArrhythmia('./data/uci_arrhythmia', train=True)
eval_dataset = dataset.UCIArrhythmia('./data/uci_arrhythmia', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'cardiotocography':
train_dataset = dataset.UCICardiotocography('./data/uci_card', train=True)
eval_dataset = dataset.UCICardiotocography('./data/uci_card', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'breastcancer':
train_dataset = dataset.UCIBreastcancer('./data/uci_breast', train=True)
eval_dataset = dataset.UCIBreastcancer('./data/uci_breast', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'nomao':
train_dataset = dataset.UCINomao('./data/uci_nomao', train=True)
eval_dataset = dataset.UCINomao('./data/uci_nomao', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'multiplefeatures':
train_dataset = dataset.UCIMultiplefeatures('./data/uci_multiple_features', train=True)
eval_dataset = dataset.UCIMultiplefeatures('./data/uci_multiple_features', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'madelon':
train_dataset = dataset.UCIMadelon('./data/uci_madelon', train=True)
eval_dataset = dataset.UCIMadelon('./data/uci_madelon', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'secom':
train_dataset = dataset.UCISecom('./data/uci_secom', train=True)
eval_dataset = dataset.UCISecom('./data/uci_secom', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
elif opt.dataset == 'isolet5':
train_dataset = dataset.UCIIsolet5('./data/uci_isolet5', train=True)
eval_dataset = dataset.UCIIsolet5('./data/uci_isolet5', train=False)
return {'train': train_dataset, 'eval': eval_dataset}
else:
raise NotImplementedError