def __init__(self, **kwargs):
"""Construct the Imagenet class."""
Dataset.__init__(self, **kwargs)
split = 'train' if self.train else 'val'
local_data_path = FileOps.join_path(self.args.data_path, split)
ImageFolder.__init__(self,
root=local_data_path,
transform=Compose(self.transforms.__transform__))
def __init__(self, root, transform=None, target_transform=None, download=False, **kwargs):
self.root = root
if download:
self.download()
if not self._check_integrity():
raise RuntimeError('Dataset not found or corrupted.' +
' You can use download=True to download it')
ImageFolder.__init__(self, os.path.join(root, self.base_folder), transform = transform, target_transform = target_transform, **kwargs)
def __init__(self,
root,
train=False,
transform=None,
target_transform=None,
download=False,
**kwargs):
self.root = root
self.transform = transform
self.target_transform = target_transform
self.loader = default_loader
if download:
download_url(self.url, self.root, self.filename, self.zip_md5)
if not self._check_integrity():
# extract file
cwd = os.getcwd()
os.chdir(root)
with zipfile.ZipFile(self.filename, "r") as zip:
zip.extractall()
os.chdir(cwd)
if not self._check_integrity():
raise RuntimeError('Dataset not found or corrupted.' +
' You can use download=True to download it')
ImageFolder.__init__(self,
os.path.join(root, self.base_folder),
transform=transform,
target_transform=target_transform,
**kwargs)
self.super_classes = self.classes
samples = []
classes = set()
f = open(
os.path.join(root, self.base_folder,
'Ebay_{}.txt'.format('train' if train else 'test')))
f.readline()
for (image_id, class_id, super_class_id, path) in map(str.split, f):
samples.append((os.path.join(root, self.base_folder,
path), int(class_id) - 1))
classes.add(
"%s.%s" %
(class_id, self.super_classes[int(super_class_id) - 1]))
self.samples = samples
self.classes = list(classes)
self.classes.sort(key=lambda x: int(x.split(".")[0]))
self.class_to_idx = {cls: i for i, cls in enumerate(self.classes)}
self.imgs = self.samples
def __init__(self, root, train=False, transform=None, target_transform=None, download=False, **kwargs):
self.root = root + "/Cars196"
self.transform = transform
self.target_transform = target_transform
self.loader = default_loader
if download:
download_url(self.img_url, self.root, self.img_filename, self.img_md5)
download_url(self.anno_url, self.root, self.anno_filename, self.anno_md5)
if not self._check_integrity():
cwd = os.getcwd()
tar = tarfile.open(os.path.join(self.root, self.img_filename), "r:gz")
os.chdir(self.root)
tar.extractall()
tar.close()
os.chdir(cwd)
if not self._check_integrity() or \
not check_integrity(os.path.join(self.root, self.anno_filename), self.anno_md5):
raise RuntimeError('Dataset not found or corrupted.' +
' You can use download=True to download it')
ImageFolder.__init__(self, os.path.join(self.root),
transform=transform, target_transform=target_transform, **kwargs)
self.root = root + "/Cars196"
labels = io.loadmat(os.path.join(self.root, self.anno_filename))['annotations'][0]
class_names = io.loadmat(os.path.join(self.root, self.anno_filename))['class_names'][0]
if train:
self.classes = [str(c[0]) for c in class_names[:98]]
self.class_to_idx = {cls: i for i, cls in enumerate(self.classes)}
else:
self.classes = [str(c[0]) for c in class_names[98:]]
self.class_to_idx = {cls: i+98 for i, cls in enumerate(self.classes)}
class_idx = list(self.class_to_idx.values())
samples = []
for l in labels:
cls = int(l[5][0, 0]) - 1
p = l[0][0]
if cls in class_idx:
samples.append((os.path.join(self.root, p), int(cls)))
self.samples = samples
self.imgs = self.samples
def __init__(
self,
root: str = default_dataset_path("imagenette"),
train: bool = True,
rand_trans: bool = False,
dataset_size: ImagenetteSize = ImagenetteSize.s160,
image_size: Union[int, None] = None,
download: bool = True,
):
if torchvision_import_error is not None:
raise torchvision_import_error
ImagenetteDownloader.__init__(self, root, dataset_size, download)
if image_size is None:
image_size = 160 if dataset_size == ImagenetteSize.s160 else 224
if rand_trans:
trans = [
transforms.RandomResizedCrop(image_size),
transforms.RandomHorizontalFlip(),
]
else:
resize_scale = 256.0 / 224.0 # standard used
trans = [
transforms.Resize(round(resize_scale * image_size)),
transforms.CenterCrop(image_size),
]
trans.extend([
transforms.ToTensor(),
transforms.Normalize(mean=IMAGENET_RGB_MEANS,
std=IMAGENET_RGB_STDS),
])
ImageFolder.__init__(self, self.split_root(train),
transforms.Compose(trans))
# make sure we don't preserve the folder structure class order
random.shuffle(self.samples)
