def get_dataloaders(dataset,
batch,
dataroot,
split=0.0,
split_idx=0,
horovod=False):
if 'cifar' in dataset:
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)),
])
elif 'imagenet' in dataset:
transform_train = transforms.Compose([
transforms.RandomResizedCrop(224),
transforms.RandomHorizontalFlip(),
transforms.ColorJitter(
brightness=0.4,
contrast=0.4,
saturation=0.4,
hue=0.2,
),
transforms.ToTensor(),
# Lighting(0.1, _IMAGENET_PCA['eigval'], _IMAGENET_PCA['eigvec']),
transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
])
if C.get()['model']['type'] == 'resnet200':
# Instead, we test a single 320×320 crop from s = 320
transform_test = transforms.Compose([
transforms.Resize(320),
transforms.CenterCrop(320),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
])
else:
transform_test = transforms.Compose([
transforms.Resize(256),
transforms.CenterCrop(224),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
])
else:
raise ValueError('dataset=%s' % dataset)
if isinstance(C.get()['aug'], list):
logger.debug('augmentation provided.')
transform_train.transforms.insert(0, Augmentation(C.get()['aug']))
else:
logger.debug('augmentation: %s' % C.get()['aug'])
if C.get()['aug'] == 'random2048':
transform_train.transforms.insert(
0, Augmentation(random_search2048()))
elif C.get()['aug'] == 'fa_reduced_cifar10':
transform_train.transforms.insert(
0, Augmentation(fa_reduced_cifar10()))
elif C.get()['aug'] == 'fa_reduced_imagenet':
transform_train.transforms.insert(
0, Augmentation(fa_reduced_imagenet()))
elif C.get()['aug'] == 'arsaug':
transform_train.transforms.insert(0, Augmentation(arsaug_policy()))
elif C.get()['aug'] == 'autoaug_cifar10':
transform_train.transforms.insert(
0, Augmentation(autoaug_paper_cifar10()))
elif C.get()['aug'] == 'autoaug_extend':
transform_train.transforms.insert(0,
Augmentation(autoaug_policy()))
elif C.get()['aug'] in ['default', 'inception', 'inception320']:
pass
else:
raise ValueError('not found augmentations. %s' % C.get()['aug'])
if C.get()['cutout'] > 0:
transform_train.transforms.append(CutoutDefault(C.get()['cutout']))
if dataset == 'cifar10':
total_trainset = torchvision.datasets.CIFAR10(
root=dataroot,
train=True,
download=True,
transform=transform_train)
testset = torchvision.datasets.CIFAR10(root=dataroot,
train=False,
download=True,
transform=transform_test)
elif dataset == 'reduced_cifar10':
total_trainset = torchvision.datasets.CIFAR10(
root=dataroot,
train=True,
download=True,
transform=transform_train)
sss = StratifiedShuffleSplit(n_splits=1,
test_size=46000,
random_state=0) # 4000 trainset
sss = sss.split(list(range(len(total_trainset))),
total_trainset.train_labels)
train_idx, valid_idx = next(sss)
train_labels = [total_trainset.train_labels[idx] for idx in train_idx]
total_trainset = Subset(total_trainset, train_idx)
total_trainset.train_labels = train_labels
testset = torchvision.datasets.CIFAR10(root=dataroot,
train=False,
download=True,
transform=transform_test)
elif dataset == 'cifar100':
total_trainset = torchvision.datasets.CIFAR100(
root=dataroot,
train=True,
download=True,
transform=transform_train)
testset = torchvision.datasets.CIFAR100(root=dataroot,
train=False,
download=True,
transform=transform_test)
elif dataset == 'imagenet':
total_trainset = torchvision.datasets.ImageFolder(
root=os.path.join(dataroot, 'imagenet/train'),
transform=transform_train)
testset = torchvision.datasets.ImageFolder(root=os.path.join(
dataroot, 'imagenet/val'),
transform=transform_test)
# compatibility
total_trainset.train_labels = [lb for _, lb in total_trainset.samples]
elif dataset == 'reduced_imagenet':
# randomly chosen indices
idx120 = [
904, 385, 759, 884, 784, 844, 132, 214, 990, 786, 979, 582, 104,
288, 697, 480, 66, 943, 308, 282, 118, 926, 882, 478, 133, 884,
570, 964, 825, 656, 661, 289, 385, 448, 705, 609, 955, 5, 703, 713,
695, 811, 958, 147, 6, 3, 59, 354, 315, 514, 741, 525, 685, 673,
657, 267, 575, 501, 30, 455, 905, 860, 355, 911, 24, 708, 346, 195,
660, 528, 330, 511, 439, 150, 988, 940, 236, 803, 741, 295, 111,
520, 856, 248, 203, 147, 625, 589, 708, 201, 712, 630, 630, 367,
273, 931, 960, 274, 112, 239, 463, 355, 955, 525, 404, 59, 981,
725, 90, 782, 604, 323, 418, 35, 95, 97, 193, 690, 869, 172
]
total_trainset = torchvision.datasets.ImageFolder(
root=os.path.join(dataroot, 'imagenet/train'),
transform=transform_train)
testset = torchvision.datasets.ImageFolder(root=os.path.join(
dataroot, 'imagenet/val'),
transform=transform_test)
# compatibility
total_trainset.train_labels = [lb for _, lb in total_trainset.samples]
sss = StratifiedShuffleSplit(n_splits=1,
test_size=len(total_trainset) - 500000,
random_state=0) # 4000 trainset
sss = sss.split(list(range(len(total_trainset))),
total_trainset.train_labels)
train_idx, valid_idx = next(sss)
# filter out
train_idx = list(
filter(lambda x: total_trainset.train_labels[x] in idx120,
train_idx))
valid_idx = list(
filter(lambda x: total_trainset.train_labels[x] in idx120,
valid_idx))
test_idx = list(
filter(lambda x: testset.samples[x][1] in idx120,
range(len(testset))))
train_labels = [
idx120.index(total_trainset.train_labels[idx]) for idx in train_idx
]
for idx in range(len(total_trainset.samples)):
if total_trainset.samples[idx][1] not in idx120:
continue
total_trainset.samples[idx] = (total_trainset.samples[idx][0],
idx120.index(
total_trainset.samples[idx][1]))
total_trainset = Subset(total_trainset, train_idx)
total_trainset.train_labels = train_labels
for idx in range(len(testset.samples)):
if testset.samples[idx][1] not in idx120:
continue
testset.samples[idx] = (testset.samples[idx][0],
idx120.index(testset.samples[idx][1]))
testset = Subset(testset, test_idx)
print('reduced_imagenet train=', len(total_trainset))
else:
raise ValueError('invalid dataset name=%s' % dataset)
if split > 0.0:
sss = StratifiedShuffleSplit(n_splits=5,
test_size=split,
random_state=0)
sss = sss.split(list(range(len(total_trainset))),
total_trainset.train_labels)
for _ in range(split_idx + 1):
train_idx, valid_idx = next(sss)
train_sampler = SubsetRandomSampler(train_idx)
valid_sampler = SubsetSampler(valid_idx)
if horovod:
import horovod.torch as hvd
train_sampler = torch.utils.data.distributed.DistributedSampler(
train_sampler, num_replicas=hvd.size(), rank=hvd.rank())
else:
valid_sampler = SubsetSampler([])
if horovod:
import horovod.torch as hvd
train_sampler = DistributedStratifiedSampler(
total_trainset.train_labels,
num_replicas=hvd.size(),
rank=hvd.rank())
else:
train_sampler = StratifiedSampler(total_trainset.train_labels)
trainloader = torch.utils.data.DataLoader(
total_trainset,
batch_size=batch,
shuffle=True if train_sampler is None else False,
num_workers=32,
pin_memory=True,
sampler=train_sampler,
drop_last=True)
validloader = torch.utils.data.DataLoader(total_trainset,
batch_size=batch,
shuffle=False,
num_workers=16,
pin_memory=True,
sampler=valid_sampler,
drop_last=False)
testloader = torch.utils.data.DataLoader(testset,
batch_size=batch,
shuffle=False,
num_workers=32,
pin_memory=True,
drop_last=False)
return train_sampler, trainloader, validloader, testloader
def load_dataset(data_dir, resize, dataset_name, img_type):
if dataset_name == 'cifar_10':
mean = cifar_10['mean']
std = cifar_10['std']
elif dataset_name == 'cifar_100':
mean = cifar_100['mean']
std = cifar_100['std']
else:
print(
'Dataset not recognized. Data normalize with equal mean/std weights'
)
mean = [0.5, 0.5, 0.5]
std = [0.5, 0.5, 0.5]
hdf5_folder = '{}/hdf5'.format(data_dir)
if os.path.exists(hdf5_folder):
shutil.rmtree(hdf5_folder)
create_hdf5(data_dir, resize, dataset_name, img_type)
train_transform = transforms.Compose([
transforms.Pad(4),
transforms.RandomCrop(resize),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize(mean=mean, std=std)
])
test_transform = transforms.Compose(
[transforms.ToTensor(),
transforms.Normalize(mean=mean, std=std)])
if isinstance(C.get()['aug'], list):
logger.debug('augmentation provided.')
train_transform.transforms.insert(0, Augmentation(C.get()['aug']))
else:
logger.debug('augmentation: %s' % C.get()['aug'])
if C.get()['aug'] == 'random2048':
train_transform.transforms.insert(
0, Augmentation(random_search2048()))
elif C.get()['aug'] == 'fa_reduced_cifar10':
train_transform.transforms.insert(
0, Augmentation(fa_reduced_cifar10()))
elif C.get()['aug'] == 'fa_reduced_imagenet':
train_transform.transforms.insert(
0, Augmentation(fa_reduced_imagenet()))
elif C.get()['aug'] == 'arsaug':
train_transform.transforms.insert(0, Augmentation(arsaug_policy()))
elif C.get()['aug'] == 'autoaug_cifar10':
train_transform.transforms.insert(
0, Augmentation(autoaug_paper_cifar10()))
elif C.get()['aug'] == 'autoaug_extend':
train_transform.transforms.insert(0,
Augmentation(autoaug_policy()))
elif C.get()['aug'] in ['default', 'inception', 'inception320']:
pass
else:
raise ValueError('not found augmentations. %s' % C.get()['aug'])
if C.get()['cutout'] > 0:
train_transform.transforms.append(CutoutDefault(C.get()['cutout']))
hdf5_folder = '{}/hdf5'.format(data_dir)
hdf5_train_path = '{}/{}_{}.hdf5'.format(hdf5_folder, dataset_name,
'training')
hdf5_test_path = '{}/{}_{}.hdf5'.format(hdf5_folder, dataset_name, 'test')
train_dataset = CustomDataset(hdf5_file=hdf5_train_path,
transform=train_transform)
val_dataset = CustomDataset(hdf5_file=hdf5_train_path,
transform=test_transform)
test_dataset = CustomDataset(hdf5_file=hdf5_test_path,
transform=test_transform)
train_dataset.train_labels = train_dataset.labels_id
return [train_dataset, val_dataset, test_dataset]
def get_dataloaders(dataset, batch, dataroot, split=0.15, split_idx=0, multinode=False, target_lb=-1, gr_assign=None, gr_id=None, gr_ids=None, rand_val=False):
if 'cifar' in dataset or 'svhn' in dataset:
if "cifar" in dataset:
_mean, _std = _CIFAR_MEAN, _CIFAR_STD
else:
_mean, _std = _SVHN_MEAN, _SVHN_STD
transform_train = transforms.Compose([
transforms.RandomCrop(32, padding=4),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize(_mean, _std),
])
transform_test = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize(_mean, _std),
])
elif 'imagenet' in dataset:
input_size = 224
sized_size = 256
if 'efficientnet' in C.get()['model']['type']:
input_size = EfficientNet.get_image_size(C.get()['model']['type'])
sized_size = input_size + 32 # TODO
# sized_size = int(round(input_size / 224. * 256))
# sized_size = input_size
logger.info('size changed to %d/%d.' % (input_size, sized_size))
transform_train = transforms.Compose([
EfficientNetRandomCrop(input_size),
transforms.Resize((input_size, input_size), interpolation=Image.BICUBIC),
# transforms.RandomResizedCrop(input_size, scale=(0.1, 1.0), interpolation=Image.BICUBIC),
transforms.RandomHorizontalFlip(),
transforms.ColorJitter(
brightness=0.4,
contrast=0.4,
saturation=0.4,
),
transforms.ToTensor(),
Lighting(0.1, _IMAGENET_PCA['eigval'], _IMAGENET_PCA['eigvec']),
transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
])
transform_test = transforms.Compose([
EfficientNetCenterCrop(input_size),
transforms.Resize((input_size, input_size), interpolation=Image.BICUBIC),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
])
else:
raise ValueError('dataset=%s' % dataset)
if isinstance(C.get()['aug'], list):
logger.debug('augmentation provided.')
transform_train.transforms.insert(0, Augmentation(C.get()['aug']))
elif isinstance(C.get()['aug'], dict):
# group version
logger.debug('group augmentation provided.')
else:
logger.debug('augmentation: %s' % C.get()['aug'])
if C.get()['aug'] == 'fa_reduced_cifar10':
transform_train.transforms.insert(0, Augmentation(fa_reduced_cifar10()))
elif C.get()['aug'] == 'fa_reduced_imagenet':
transform_train.transforms.insert(0, Augmentation(fa_resnet50_rimagenet()))
elif C.get()['aug'] == 'fa_reduced_svhn':
transform_train.transforms.insert(0, Augmentation(fa_reduced_svhn()))
elif C.get()['aug'] == 'arsaug':
transform_train.transforms.insert(0, Augmentation(arsaug_policy()))
elif C.get()['aug'] == 'autoaug_cifar10':
transform_train.transforms.insert(0, Augmentation(autoaug_paper_cifar10()))
elif C.get()['aug'] == 'autoaug_extend':
transform_train.transforms.insert(0, Augmentation(autoaug_policy()))
elif C.get()['aug'] in ['default', "clean", "nonorm", "nocut"]:
pass
else:
raise ValueError('not found augmentations. %s' % C.get()['aug'])
if C.get()['cutout'] > 0 and C.get()['aug'] != "nocut":
transform_train.transforms.append(CutoutDefault(C.get()['cutout']))
if C.get()['aug'] == "clean":
transform_train = transform_test
elif C.get()['aug'] == "nonorm":
transform_train = transforms.Compose([
transforms.ToTensor()
])
train_idx = valid_idx = None
if dataset == 'cifar10':
if isinstance(C.get()['aug'], dict):
total_trainset = GrAugCIFAR10(root=dataroot, gr_assign=gr_assign, gr_policies=C.get()['aug'], train=True, download=False, transform=transform_train)
else:
total_trainset = torchvision.datasets.CIFAR10(root=dataroot, train=True, download=False, transform=transform_train)
testset = torchvision.datasets.CIFAR10(root=dataroot, train=False, download=False, transform=transform_test)
elif dataset == 'reduced_cifar10':
if isinstance(C.get()['aug'], dict):
total_trainset = GrAugCIFAR10(root=dataroot, gr_assign=gr_assign, gr_policies=C.get()['aug'], train=True, download=False, transform=transform_train)
else:
total_trainset = torchvision.datasets.CIFAR10(root=dataroot, train=True, download=False, transform=transform_train)
sss = StratifiedShuffleSplit(n_splits=5, train_size=4000, random_state=0) # 4000 trainset
sss = sss.split(list(range(len(total_trainset))), total_trainset.targets)
for _ in range(split_idx+1):
train_idx, valid_idx = next(sss)
testset = torchvision.datasets.CIFAR10(root=dataroot, train=False, download=False, transform=transform_test)
elif dataset == 'cifar100':
if isinstance(C.get()['aug'], dict):
total_trainset = GrAugData("CIFAR100", root=dataroot, gr_assign=gr_assign, gr_policies=C.get()['aug'], train=True, download=False, transform=transform_train)
else:
total_trainset = torchvision.datasets.CIFAR100(root=dataroot, train=True, download=False, transform=transform_train)
testset = torchvision.datasets.CIFAR100(root=dataroot, train=False, download=False, transform=transform_test)
elif dataset == 'svhn': #TODO
trainset = torchvision.datasets.SVHN(root=dataroot, split='train', download=False, transform=transform_train)
extraset = torchvision.datasets.SVHN(root=dataroot, split='extra', download=False, transform=transform_train)
total_trainset = ConcatDataset([trainset, extraset])
testset = torchvision.datasets.SVHN(root=dataroot, split='test', download=False, transform=transform_test)
elif dataset == 'reduced_svhn':
if isinstance(C.get()['aug'], dict):
total_trainset = GrAugData("SVHN", root=dataroot, gr_assign=gr_assign, gr_policies=C.get()['aug'], split='train', download=False, transform=transform_train)
else:
total_trainset = torchvision.datasets.SVHN(root=dataroot, split='train', download=False, transform=transform_train)
sss = StratifiedShuffleSplit(n_splits=5, train_size=1000, test_size=7325, random_state=0)
sss = sss.split(list(range(len(total_trainset))), total_trainset.labels)
for _ in range(split_idx+1):
train_idx, valid_idx = next(sss)
# targets = [total_trainset.labels[idx] for idx in train_idx]
# total_trainset = Subset(total_trainset, train_idx)
# total_trainset.targets = targets
testset = torchvision.datasets.SVHN(root=dataroot, split='test', download=False, transform=transform_test)
elif dataset == 'imagenet':
total_trainset = ImageNet(root=os.path.join(dataroot, 'imagenet-pytorch'), transform=transform_train)
testset = ImageNet(root=os.path.join(dataroot, 'imagenet-pytorch'), split='val', transform=transform_test)
# compatibility
total_trainset.targets = [lb for _, lb in total_trainset.samples]
elif dataset == 'reduced_imagenet':
# randomly chosen indices
# idx120 = sorted(random.sample(list(range(1000)), k=120))
idx120 = [16, 23, 52, 57, 76, 93, 95, 96, 99, 121, 122, 128, 148, 172, 181, 189, 202, 210, 232, 238, 257, 258, 259, 277, 283, 289, 295, 304, 307, 318, 322, 331, 337, 338, 345, 350, 361, 375, 376, 381, 388, 399, 401, 408, 424, 431, 432, 440, 447, 462, 464, 472, 483, 497, 506, 512, 530, 541, 553, 554, 557, 564, 570, 584, 612, 614, 619, 626, 631, 632, 650, 657, 658, 660, 674, 675, 680, 682, 691, 695, 699, 711, 734, 736, 741, 754, 757, 764, 769, 770, 780, 781, 787, 797, 799, 811, 822, 829, 830, 835, 837, 842, 843, 845, 873, 883, 897, 900, 902, 905, 913, 920, 925, 937, 938, 940, 941, 944, 949, 959]
total_trainset = ImageNet(root=os.path.join(dataroot, 'imagenet-pytorch'), transform=transform_train)
testset = ImageNet(root=os.path.join(dataroot, 'imagenet-pytorch'), split='val', transform=transform_test)
# compatibility
total_trainset.targets = [lb for _, lb in total_trainset.samples]
sss = StratifiedShuffleSplit(n_splits=1, test_size=len(total_trainset) - 50000, random_state=0) # 4000 trainset
sss = sss.split(list(range(len(total_trainset))), total_trainset.targets)
train_idx, valid_idx = next(sss)
# filter out
train_idx = list(filter(lambda x: total_trainset.labels[x] in idx120, train_idx))
valid_idx = list(filter(lambda x: total_trainset.labels[x] in idx120, valid_idx))
test_idx = list(filter(lambda x: testset.samples[x][1] in idx120, range(len(testset))))
targets = [idx120.index(total_trainset.targets[idx]) for idx in train_idx]
for idx in range(len(total_trainset.samples)):
if total_trainset.samples[idx][1] not in idx120:
continue
total_trainset.samples[idx] = (total_trainset.samples[idx][0], idx120.index(total_trainset.samples[idx][1]))
total_trainset = Subset(total_trainset, train_idx)
total_trainset.targets = targets
for idx in range(len(testset.samples)):
if testset.samples[idx][1] not in idx120:
continue
testset.samples[idx] = (testset.samples[idx][0], idx120.index(testset.samples[idx][1]))
testset = Subset(testset, test_idx)
print('reduced_imagenet train=', len(total_trainset))
elif dataset == "cifar10_svhn":
if isinstance(C.get()['aug'], dict):
# last stage: benchmark test
total_trainset = GrAugMix(dataset.split("_"), gr_assign=gr_assign, gr_policies=C.get()['aug'], root=dataroot, train=True, download=False, transform=transform_train, gr_ids=gr_ids)
else:
# eval_tta & childnet training
total_trainset = GrAugMix(dataset.split("_"), root=dataroot, train=True, download=False, transform=transform_train)
testset = GrAugMix(dataset.split("_"), root=dataroot, train=False, download=False, transform=transform_test)
else:
raise ValueError('invalid dataset name=%s' % dataset)
if not hasattr(total_trainset, "gr_ids"):
total_trainset.gr_ids = None
if gr_ids is not None:
total_trainset.gr_ids = gr_ids
if gr_assign is not None and total_trainset.gr_ids is None:
# eval_tta3
temp_trainset = copy.deepcopy(total_trainset)
# temp_trainset.transform = transform_test # just normalize
temp_loader = torch.utils.data.DataLoader(
temp_trainset, batch_size=batch, shuffle=False, num_workers=4,
drop_last=False)
gr_dist = gr_assign(temp_loader)
gr_ids = torch.max(gr_dist)[1].numpy()
if split > 0.0:
if train_idx is None or valid_idx is None:
# filter by split ratio
sss = StratifiedShuffleSplit(n_splits=5, test_size=split, random_state=0)
sss = sss.split(list(range(len(total_trainset))), total_trainset.targets)
for _ in range(split_idx + 1):
train_idx, valid_idx = next(sss)
if gr_id is not None:
# filter by group
idx2gr = total_trainset.gr_ids
ps = PredefinedSplit(idx2gr)
ps = ps.split()
for _ in range(gr_id + 1):
_, gr_split_idx = next(ps)
train_idx = [idx for idx in train_idx if idx in gr_split_idx]
valid_idx = [idx for idx in valid_idx if idx in gr_split_idx]
if target_lb >= 0:
train_idx = [i for i in train_idx if total_trainset.targets[i] == target_lb]
valid_idx = [i for i in valid_idx if total_trainset.targets[i] == target_lb]
train_sampler = SubsetRandomSampler(train_idx)
valid_sampler = SubsetSampler(valid_idx) if not rand_val else SubsetRandomSampler(valid_idx)
if multinode:
train_sampler = torch.utils.data.distributed.DistributedSampler(Subset(total_trainset, train_idx), num_replicas=dist.get_world_size(), rank=dist.get_rank())
else:
train_sampler = None
valid_sampler = SubsetSampler([])
if gr_id is not None:
# filter by group
idx2gr = total_trainset.gr_ids
ps = PredefinedSplit(idx2gr)
ps = ps.split()
for _ in range(gr_id + 1):
_, gr_split_idx = next(ps)
targets = [total_trainset.targets[idx] for idx in gr_split_idx]
total_trainset = Subset(total_trainset, gr_split_idx)
total_trainset.targets = targets
if train_idx is not None and valid_idx is not None:
if dataset in ["svhn", "reduced_svhn"]:
targets = [total_trainset.labels[idx] for idx in train_idx]
else:
targets = [total_trainset.targets[idx] for idx in train_idx]
total_trainset = Subset(total_trainset, train_idx)
total_trainset.targets = targets
if multinode:
train_sampler = torch.utils.data.distributed.DistributedSampler(total_trainset, num_replicas=dist.get_world_size(), rank=dist.get_rank())
logger.info(f'----- dataset with DistributedSampler {dist.get_rank()}/{dist.get_world_size()}')
trainloader = torch.utils.data.DataLoader(
total_trainset, batch_size=batch, shuffle=True if train_sampler is None else False, num_workers=8 if torch.cuda.device_count()==8 else 4, pin_memory=True,
sampler=train_sampler, drop_last=True)
validloader = torch.utils.data.DataLoader(
total_trainset, batch_size=batch, shuffle=False, num_workers=4, pin_memory=True,
sampler=valid_sampler, drop_last=False if not rand_val else True)
testloader = torch.utils.data.DataLoader(
testset, batch_size=batch, shuffle=False, num_workers=8 if torch.cuda.device_count()==8 else 4, pin_memory=True,
drop_last=False
)
return train_sampler, trainloader, validloader, testloader
def get_dataloaders(dataset,
batch,
dataroot,
split=0.15,
split_idx=0,
multinode=False,
target_lb=-1):
if 'cifar' in dataset or 'svhn' in dataset:
transform_train = transforms.Compose([
transforms.RandomCrop(32, padding=4),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize(_CIFAR_MEAN, _CIFAR_STD),
])
transform_test = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize(_CIFAR_MEAN, _CIFAR_STD),
])
elif 'imagenet' in dataset:
input_size = 224
sized_size = 256
if 'efficientnet' in C.get()['model']['type']:
input_size = EfficientNet.get_image_size(C.get()['model']['type'])
sized_size = input_size + 32 # TODO
# sized_size = int(round(input_size / 224. * 256))
# sized_size = input_size
logger.info('size changed to %d/%d.' % (input_size, sized_size))
transform_train = transforms.Compose([
EfficientNetRandomCrop(input_size),
transforms.Resize((input_size, input_size),
interpolation=Image.BICUBIC),
# transforms.RandomResizedCrop(input_size, scale=(0.1, 1.0), interpolation=Image.BICUBIC),
transforms.RandomHorizontalFlip(),
transforms.ColorJitter(
brightness=0.4,
contrast=0.4,
saturation=0.4,
),
transforms.ToTensor(),
Lighting(0.1, _IMAGENET_PCA['eigval'], _IMAGENET_PCA['eigvec']),
transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
])
transform_test = transforms.Compose([
EfficientNetCenterCrop(input_size),
transforms.Resize((input_size, input_size),
interpolation=Image.BICUBIC),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
])
elif 'gta5' in dataset:
transform_target_after = transforms.Compose([PILToLongTensor()])
transform_train_pre = Compose(
[RandomCrop((321, 321)),
RandomHorizontallyFlip()]) # weak transform
transform_valid_pre = Compose([RandomCrop(
(321, 321))]) # weak transform
transform_train = transforms.Compose([transforms.ToTensor()])
transform_test_pre = None # Compose([RandomCrop((321, 321))])
transform_test = transforms.Compose([transforms.ToTensor()])
else:
raise ValueError('dataset=%s' % dataset)
total_aug = augs = None
if isinstance(C.get()['aug'], list):
logger.debug('augmentation provided.')
transform_train.transforms.insert(0, Augmentation(C.get()['aug']))
else:
logger.debug('augmentation: %s' % C.get()['aug'])
if C.get()['aug'] == 'fa_reduced_cifar10':
transform_train.transforms.insert(
0, Augmentation(fa_reduced_cifar10()))
elif C.get()['aug'] == 'fa_reduced_imagenet':
transform_train.transforms.insert(
0, Augmentation(fa_resnet50_rimagenet()))
elif C.get()['aug'] == 'fa_reduced_svhn':
transform_train.transforms.insert(0,
Augmentation(fa_reduced_svhn()))
elif C.get()['aug'] == 'arsaug':
transform_train.transforms.insert(0, Augmentation(arsaug_policy()))
elif C.get()['aug'] == 'autoaug_cifar10':
transform_train.transforms.insert(
0, Augmentation(autoaug_paper_cifar10()))
elif C.get()['aug'] == 'autoaug_extend':
transform_train.transforms.insert(0,
Augmentation(autoaug_policy()))
elif C.get()['aug'] in ['default']:
pass
else:
raise ValueError('not found augmentations. %s' % C.get()['aug'])
if C.get()['cutout'] > 0:
transform_train.transforms.append(CutoutDefault(C.get()['cutout']))
if dataset == 'cifar10':
total_trainset = torchvision.datasets.CIFAR10(
root=dataroot,
train=True,
download=True,
transform=transform_train)
testset = torchvision.datasets.CIFAR10(root=dataroot,
train=False,
download=True,
transform=transform_test)
elif dataset == 'reduced_cifar10':
total_trainset = torchvision.datasets.CIFAR10(
root=dataroot,
train=True,
download=True,
transform=transform_train)
sss = StratifiedShuffleSplit(n_splits=1,
test_size=46000,
random_state=0) # 4000 trainset
sss = sss.split(list(range(len(total_trainset))),
total_trainset.targets)
train_idx, valid_idx = next(sss)
targets = [total_trainset.targets[idx] for idx in train_idx]
total_trainset = Subset(total_trainset, train_idx)
total_trainset.targets = targets
testset = torchvision.datasets.CIFAR10(root=dataroot,
train=False,
download=True,
transform=transform_test)
elif dataset == 'cifar100':
total_trainset = torchvision.datasets.CIFAR100(
root=dataroot,
train=True,
download=True,
transform=transform_train)
testset = torchvision.datasets.CIFAR100(root=dataroot,
train=False,
download=True,
transform=transform_test)
elif dataset == 'svhn':
trainset = torchvision.datasets.SVHN(root=dataroot,
split='train',
download=True,
transform=transform_train)
extraset = torchvision.datasets.SVHN(root=dataroot,
split='extra',
download=True,
transform=transform_train)
total_trainset = ConcatDataset([trainset, extraset])
testset = torchvision.datasets.SVHN(root=dataroot,
split='test',
download=True,
transform=transform_test)
elif dataset == 'reduced_svhn':
total_trainset = torchvision.datasets.SVHN(root=dataroot,
split='train',
download=True,
transform=transform_train)
sss = StratifiedShuffleSplit(n_splits=1,
test_size=73257 - 1000,
random_state=0) # 1000 trainset
sss = sss.split(list(range(len(total_trainset))),
total_trainset.targets)
train_idx, valid_idx = next(sss)
targets = [total_trainset.targets[idx] for idx in train_idx]
total_trainset = Subset(total_trainset, train_idx)
total_trainset.targets = targets
testset = torchvision.datasets.SVHN(root=dataroot,
split='test',
download=True,
transform=transform_test)
elif dataset == 'imagenet':
total_trainset = ImageNet(root=os.path.join(dataroot,
'imagenet-pytorch'),
transform=transform_train,
download=True)
testset = ImageNet(root=os.path.join(dataroot, 'imagenet-pytorch'),
split='val',
transform=transform_test)
# compatibility
total_trainset.targets = [lb for _, lb in total_trainset.samples]
elif dataset == 'reduced_imagenet':
# randomly chosen indices
# idx120 = sorted(random.sample(list(range(1000)), k=120))
idx120 = [
16, 23, 52, 57, 76, 93, 95, 96, 99, 121, 122, 128, 148, 172, 181,
189, 202, 210, 232, 238, 257, 258, 259, 277, 283, 289, 295, 304,
307, 318, 322, 331, 337, 338, 345, 350, 361, 375, 376, 381, 388,
399, 401, 408, 424, 431, 432, 440, 447, 462, 464, 472, 483, 497,
506, 512, 530, 541, 553, 554, 557, 564, 570, 584, 612, 614, 619,
626, 631, 632, 650, 657, 658, 660, 674, 675, 680, 682, 691, 695,
699, 711, 734, 736, 741, 754, 757, 764, 769, 770, 780, 781, 787,
797, 799, 811, 822, 829, 830, 835, 837, 842, 843, 845, 873, 883,
897, 900, 902, 905, 913, 920, 925, 937, 938, 940, 941, 944, 949,
959
]
total_trainset = ImageNet(root=os.path.join(dataroot,
'imagenet-pytorch'),
transform=transform_train)
testset = ImageNet(root=os.path.join(dataroot, 'imagenet-pytorch'),
split='val',
transform=transform_test)
# compatibility
total_trainset.targets = [lb for _, lb in total_trainset.samples]
sss = StratifiedShuffleSplit(n_splits=1,
test_size=len(total_trainset) - 50000,
random_state=0) # 4000 trainset
sss = sss.split(list(range(len(total_trainset))),
total_trainset.targets)
train_idx, valid_idx = next(sss)
# filter out
train_idx = list(
filter(lambda x: total_trainset.labels[x] in idx120, train_idx))
valid_idx = list(
filter(lambda x: total_trainset.labels[x] in idx120, valid_idx))
test_idx = list(
filter(lambda x: testset.samples[x][1] in idx120,
range(len(testset))))
targets = [
idx120.index(total_trainset.targets[idx]) for idx in train_idx
]
for idx in range(len(total_trainset.samples)):
if total_trainset.samples[idx][1] not in idx120:
continue
total_trainset.samples[idx] = (total_trainset.samples[idx][0],
idx120.index(
total_trainset.samples[idx][1]))
total_trainset = Subset(total_trainset, train_idx)
total_trainset.targets = targets
for idx in range(len(testset.samples)):
if testset.samples[idx][1] not in idx120:
continue
testset.samples[idx] = (testset.samples[idx][0],
idx120.index(testset.samples[idx][1]))
testset = Subset(testset, test_idx)
print('reduced_imagenet train=', len(total_trainset))
elif dataset == 'gta5':
total_trainset = GTA5_Dataset(
data_root_path=dataroot,
split='train',
transform_pre=transform_train_pre,
transform_target_after=transform_target_after,
transform_after=transform_train,
sample=10000)
total_validset = GTA5_Dataset(
data_root_path=dataroot,
split='valid',
transform_pre=transform_valid_pre,
transform_target_after=transform_target_after,
transform_after=transform_test,
sample=5000,
seed=0)
testset = GTA5_Dataset(data_root_path=dataroot,
split='valid',
transform_pre=transform_test_pre,
transform_target_after=transform_target_after,
transform_after=transform_test)
elif dataset == 'reduced_gta5':
total_trainset = GTA5_Dataset(
data_root_path=dataroot,
split='train',
transform_pre=transform_train_pre,
transform_target_after=transform_target_after,
transform_after=transform_train,
sample=1000,
seed=0)
total_validset = GTA5_Dataset(
data_root_path=dataroot,
split='valid',
transform_pre=transform_valid_pre,
transform_target_after=transform_target_after,
transform_after=transform_test,
sample=1000,
seed=0)
testset = GTA5_Dataset(data_root_path=dataroot,
split='valid',
transform_pre=transform_test_pre,
transform_target_after=transform_target_after,
transform_after=transform_test)
else:
raise ValueError('invalid dataset name=%s' % dataset)
if total_aug is not None and augs is not None:
total_trainset.set_preaug(augs, total_aug)
print('set_preaug-')
train_sampler = None
if split > 0.0:
if 'gta' not in dataset:
sss = StratifiedShuffleSplit(n_splits=5,
test_size=split,
random_state=0)
sss = sss.split(list(range(len(total_trainset))),
total_trainset.targets)
for _ in range(split_idx + 1):
train_idx, valid_idx = next(sss)
if target_lb >= 0:
train_idx = [
i for i in train_idx
if total_trainset.targets[i] == target_lb
]
valid_idx = [
i for i in valid_idx
if total_trainset.targets[i] == target_lb
]
train_sampler = SubsetRandomSampler(train_idx)
valid_sampler = SubsetSampler(valid_idx)
if multinode:
train_sampler = torch.utils.data.distributed.DistributedSampler(
Subset(total_trainset, train_idx),
num_replicas=dist.get_world_size(),
rank=dist.get_rank())
else:
train_sampler = None
valid_sampler = None
else:
valid_sampler = SubsetSampler([])
if multinode:
train_sampler = torch.utils.data.distributed.DistributedSampler(
total_trainset,
num_replicas=dist.get_world_size(),
rank=dist.get_rank())
logger.info(
f'----- dataset with DistributedSampler {dist.get_rank()}/{dist.get_world_size()}'
)
trainloader = torch.utils.data.DataLoader(
total_trainset,
batch_size=batch,
shuffle=True if train_sampler is None else False,
num_workers=8,
pin_memory=True,
sampler=train_sampler,
drop_last=True)
validloader = torch.utils.data.DataLoader(total_trainset,
batch_size=batch,
shuffle=False,
num_workers=4,
pin_memory=True,
sampler=valid_sampler,
drop_last=False)
testloader = torch.utils.data.DataLoader(testset,
batch_size=batch,
shuffle=False,
num_workers=8,
pin_memory=True,
drop_last=False)
return train_sampler, trainloader, validloader, testloader
def get_dataloaders(dataset,
batch,
dataroot,
split=0.15,
split_idx=0,
horovod=False,
target_lb=-1):
# torchvision 0.2(sh36 r0.3.0): train_labels
# torchvision 0.4(local): targets
# torchvision 0.4.1(sh36 r0.3.2): (not have attr '__version__')targets
using_attr_train_labels = False
try:
torchvision_version = torchvision.__version__
if torchvision_version < '0.4':
using_attr_train_labels = True
except AttributeError:
pass
if 'cifar' in dataset or 'svhn' in dataset:
transform_train = transforms.Compose([
transforms.RandomCrop(32, padding=4),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize(_CIFAR_MEAN, _CIFAR_STD),
])
transform_test = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize(_CIFAR_MEAN, _CIFAR_STD),
])
elif 'imagenet' in dataset:
transform_train = transforms.Compose([
transforms.RandomResizedCrop(224,
scale=(0.08, 1.0),
interpolation=Image.BICUBIC),
transforms.RandomHorizontalFlip(),
transforms.ColorJitter(
brightness=0.4,
contrast=0.4,
saturation=0.4,
),
transforms.ToTensor(),
Lighting(0.1, _IMAGENET_PCA['eigval'], _IMAGENET_PCA['eigvec']),
transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
])
transform_test = transforms.Compose([
transforms.Resize(256, interpolation=Image.BICUBIC),
transforms.CenterCrop(224),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
])
else:
raise ValueError('dataset=%s' % dataset)
total_aug = augs = None
if isinstance(C.get()['aug'], list):
logger.debug('augmentation provided.')
transform_train.transforms.insert(0, Augmentation(C.get()['aug']))
else:
logger.debug('augmentation: %s' % C.get()['aug'])
if C.get()['aug'] == 'fa_reduced_cifar10':
transform_train.transforms.insert(
0, Augmentation(fa_reduced_cifar10()))
elif C.get()['aug'] == 'fa_reduced_imagenet':
transform_train.transforms.insert(
0, Augmentation(fa_resnet50_rimagenet()))
elif C.get()['aug'] == 'fa_reduced_svhn':
transform_train.transforms.insert(0,
Augmentation(fa_reduced_svhn()))
elif C.get()['aug'] == 'arsaug':
transform_train.transforms.insert(0, Augmentation(arsaug_policy()))
elif C.get()['aug'] == 'autoaug_cifar10':
transform_train.transforms.insert(
0, Augmentation(autoaug_paper_cifar10()))
elif C.get()['aug'] == 'autoaug_extend':
transform_train.transforms.insert(0,
Augmentation(autoaug_policy()))
elif C.get()['aug'] in ['default', 'inception', 'inception320']:
pass
else:
raise ValueError('not found augmentations. %s' % C.get()['aug'])
if C.get()['cutout'] > 0:
transform_train.transforms.append(CutoutDefault(C.get()['cutout']))
if dataset == 'cifar10':
total_trainset = torchvision.datasets.CIFAR10(
root=dataroot,
train=True,
download=False,
transform=transform_train)
testset = torchvision.datasets.CIFAR10(root=dataroot,
train=False,
download=False,
transform=transform_test)
elif dataset == 'reduced_cifar10':
total_trainset = torchvision.datasets.CIFAR10(
root=dataroot,
train=True,
download=False,
transform=transform_train)
sss = StratifiedShuffleSplit(n_splits=1,
test_size=46000,
random_state=0) # 4000 trainset
sss = sss.split(list(range(len(total_trainset))),
total_trainset.targets)
train_idx, valid_idx = next(sss)
targets = [total_trainset.targets[idx] for idx in train_idx]
total_trainset = Subset(total_trainset, train_idx)
total_trainset.targets = targets
testset = torchvision.datasets.CIFAR10(root=dataroot,
train=False,
download=False,
transform=transform_test)
elif dataset == 'cifar100':
total_trainset = torchvision.datasets.CIFAR100(
root=dataroot,
train=True,
download=False,
transform=transform_train)
testset = torchvision.datasets.CIFAR100(root=dataroot,
train=False,
download=False,
transform=transform_test)
elif dataset == 'svhn':
trainset = torchvision.datasets.SVHN(root=dataroot,
split='train',
download=False,
transform=transform_train)
extraset = torchvision.datasets.SVHN(root=dataroot,
split='extra',
download=False,
transform=transform_train)
total_trainset = ConcatDataset([trainset, extraset])
testset = torchvision.datasets.SVHN(root=dataroot,
split='test',
download=False,
transform=transform_test)
elif dataset == 'reduced_svhn':
total_trainset = torchvision.datasets.SVHN(root=dataroot,
split='train',
download=False,
transform=transform_train)
sss = StratifiedShuffleSplit(n_splits=1,
test_size=73257 - 1000,
random_state=0) # 1000 trainset
sss = sss.split(
list(range(len(total_trainset))), total_trainset.train_labels
if using_attr_train_labels else total_trainset.targets)
train_idx, valid_idx = next(sss)
targets = [total_trainset.targets[idx] for idx in train_idx]
total_trainset = Subset(total_trainset, train_idx)
total_trainset.targets = targets
testset = torchvision.datasets.SVHN(root=dataroot,
split='test',
download=False,
transform=transform_test)
elif dataset == 'imagenet':
total_trainset = ImageNet(root=os.path.join(dataroot,
'imagenet-pytorch'),
transform=transform_train)
testset = ImageNet(root=os.path.join(dataroot, 'imagenet-pytorch'),
split='val',
transform=transform_test)
# compatibility
total_trainset.targets = [lb for _, lb in total_trainset.samples]
elif dataset == 'reduced_imagenet':
# randomly chosen indices
idx120 = [
904, 385, 759, 884, 784, 844, 132, 214, 990, 786, 979, 582, 104,
288, 697, 480, 66, 943, 308, 282, 118, 926, 882, 478, 133, 884,
570, 964, 825, 656, 661, 289, 385, 448, 705, 609, 955, 5, 703, 713,
695, 811, 958, 147, 6, 3, 59, 354, 315, 514, 741, 525, 685, 673,
657, 267, 575, 501, 30, 455, 905, 860, 355, 911, 24, 708, 346, 195,
660, 528, 330, 511, 439, 150, 988, 940, 236, 803, 741, 295, 111,
520, 856, 248, 203, 147, 625, 589, 708, 201, 712, 630, 630, 367,
273, 931, 960, 274, 112, 239, 463, 355, 955, 525, 404, 59, 981,
725, 90, 782, 604, 323, 418, 35, 95, 97, 193, 690, 869, 172
]
total_trainset = ImageNet(root=os.path.join(dataroot,
'imagenet-pytorch'),
transform=transform_train)
testset = ImageNet(root=os.path.join(dataroot, 'imagenet-pytorch'),
split='val',
transform=transform_test)
# compatibility
total_trainset.targets = [lb for _, lb in total_trainset.samples]
sss = StratifiedShuffleSplit(n_splits=1,
test_size=len(total_trainset) - 500000,
random_state=0) # 4000 trainset
sss = sss.split(list(range(len(total_trainset))),
total_trainset.targets)
train_idx, valid_idx = next(sss)
# filter out
train_idx = list(
filter(lambda x: total_trainset.labels[x] in idx120, train_idx))
valid_idx = list(
filter(lambda x: total_trainset.labels[x] in idx120, valid_idx))
test_idx = list(
filter(lambda x: testset.samples[x][1] in idx120,
range(len(testset))))
targets = [
idx120.index(total_trainset.targets[idx]) for idx in train_idx
]
for idx in range(len(total_trainset.samples)):
if total_trainset.samples[idx][1] not in idx120:
continue
total_trainset.samples[idx] = (total_trainset.samples[idx][0],
idx120.index(
total_trainset.samples[idx][1]))
total_trainset = Subset(total_trainset, train_idx)
total_trainset.targets = targets
for idx in range(len(testset.samples)):
if testset.samples[idx][1] not in idx120:
continue
testset.samples[idx] = (testset.samples[idx][0],
idx120.index(testset.samples[idx][1]))
testset = Subset(testset, test_idx)
print('reduced_imagenet train=', len(total_trainset))
else:
raise ValueError('invalid dataset name=%s' % dataset)
if total_aug is not None and augs is not None:
total_trainset.set_preaug(augs, total_aug)
print('set_preaug-')
train_sampler = None
if split > 0.0:
sss = StratifiedShuffleSplit(n_splits=5,
test_size=split,
random_state=0)
sss = sss.split(
list(range(len(total_trainset))), total_trainset.train_labels
if using_attr_train_labels else total_trainset.targets)
for _ in range(split_idx + 1):
train_idx, valid_idx = next(sss)
if target_lb >= 0:
train_idx = [
i for i in train_idx if total_trainset.targets[i] == target_lb
]
valid_idx = [
i for i in valid_idx if total_trainset.targets[i] == target_lb
]
train_sampler = SubsetRandomSampler(train_idx)
valid_sampler = SubsetSampler(valid_idx)
if horovod:
import horovod.torch as hvd
train_sampler = torch.utils.data.distributed.DistributedSampler(
train_sampler, num_replicas=hvd.size(), rank=hvd.rank())
else:
valid_sampler = SubsetSampler([])
if horovod:
import horovod.torch as hvd
train_sampler = torch.utils.data.distributed.DistributedSampler(
valid_sampler, num_replicas=hvd.size(), rank=hvd.rank())
trainloader = torch.utils.data.DataLoader(
total_trainset,
batch_size=batch,
shuffle=True if train_sampler is None else False,
num_workers=32,
pin_memory=True,
sampler=train_sampler,
drop_last=True)
validloader = torch.utils.data.DataLoader(total_trainset,
batch_size=batch,
shuffle=False,
num_workers=16,
pin_memory=True,
sampler=valid_sampler,
drop_last=False)
testloader = torch.utils.data.DataLoader(testset,
batch_size=batch,
shuffle=False,
num_workers=32,
pin_memory=True,
drop_last=False)
return train_sampler, trainloader, validloader, testloader