transform_train = transforms.Compose([
transforms.RandomCrop(32, padding=4),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize((0.4914, 0.4822, 0.4465),
(0.2023, 0.1994, 0.2010)),
])
transform_test = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.4914, 0.4822, 0.4465),
(0.2023, 0.1994, 0.2010)),
])
trainset = data_providers.CIFAR10(root='data',
set_name='train',
download=True,
transform=transform_train)
train_data = torch.utils.data.DataLoader(trainset,
batch_size=100,
shuffle=True,
num_workers=2)
valset = data_providers.CIFAR10(root='data',
set_name='val',
download=True,
transform=transform_test)
val_data = torch.utils.data.DataLoader(valset,
batch_size=100,
shuffle=False,
num_workers=2)
#trial - no transforms
# transform_train = transforms.Compose([
# transforms.RandomCrop(32, padding=4),
# transforms.RandomHorizontalFlip(),
# transforms.ToTensor(),
# transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010)),
# ])
# transform_test = transforms.Compose([
# transforms.ToTensor(),
# transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010)),
# ])
trainset = data_providers.CIFAR10(root='CIFAR10_Data/',
set_name='train',
download=False,
transform=transform_train)
styled = data_providers.AugmentedCIFAR10(path='CIFAR10_Data/Pickled/',
transform=transform_train)
concatenated = torch.utils.data.ConcatDataset([trainset, styled])
train_data = torch.utils.data.DataLoader(concatenated,
batch_size=128,
shuffle=True,
num_workers=4)
print('Total elements for training:', len(concatenated), len(styled))
print(styled[9][1].dtype)
print(trainset[9][1].dtype)