def __init__(self,
positive=True,
limit_sentences=None,
dataset_cache_dir=None,
dataset_name=None):
Dataset.__init__(self,
limit_sentences=limit_sentences,
dataset_cache_dir=dataset_cache_dir,
dataset_name=dataset_name)
self.positive = positive
def __init__(self, args, root, split, labelpath, cachedir, transform=None, target_transform=None, test_gap=50):
Dataset.__init__(self, test_gap, split)
self.num_classes = 174
self.transform = transform
self.target_transform = target_transform
self.cls2int = self.parse_something_labels(args.label_file)
self.labels = self.parse_something_json(labelpath, self.cls2int)
self.root = root
cachename = '{}/{}_{}.pkl'.format(cachedir, self.__class__.__name__, split)
self._data = cache(cachename)(self._prepare)(root, self.labels, split)
def __init__(self, txt_fn="data/train_phototourism_ms.txt", style2style=False):
Dataset.__init__(self)
self.txt_fn = txt_fn
self.style2style = style2style
self.path_orig_img = "/ssd/data/phototourism/orig"
self.path_style_img = "/data/datasets/phototourism/style_transfer_all"
self.style2fnames = get_style2fnames()
# a dirty hack
self.crop_size = 192
self.imgs = [line.rstrip("\n") for line in open(self.txt_fn)]
self.nimg = len(self.imgs)
self.npairs = len(self.imgs)
def __init__(self,
args,
root,
split,
labelpath,
cachedir,
transform=None,
target_transform=None,
input_size=224,
test_gap=10):
Dataset.__init__(self, test_gap, split)
self.num_classes = 80
self.transform = transform
self.target_transform = target_transform
self.cls2int = dict((str(x + 1), x) for x in range(80))
self.labels = self.parse_ava_csv(labelpath, self.cls2int)
self.root = root
self.train_gap = 64
self.input_size = input_size
cachename = '{}/{}_{}.pkl'.format(cachedir, self.__class__.__name__,
split)
self._data = cache(cachename)(self._prepare)(root, self.labels, split)
def __init__(self,
args,
root,
split,
label_path,
cachedir,
transform=None,
target_transform=None,
input_size=224,
test_gap=25,
train_gap=4):
Dataset.__init__(self, test_gap, split)
self.num_classes = 400
self.transform = transform
self.target_transform = target_transform
self.cls2int = self.parse_kinetics_labels(args.train_file)
self.labels = self.parse_kinetics_csv(label_path, self.cls2int)
self.root = root
self.train_gap = train_gap
self.input_size = input_size
cachename = '{}/{}_{}.pkl'.format(cachedir, self.__class__.__name__,
split)
self._data = cache(cachename)(self._prepare)(root, self.labels, split)
def __init__(self):
Dataset.__init__(self,
data=ISBSG10.data(),
dec_meta=ISBSG10.decision_meta(),
obj_meta=ISBSG10.objective_meta())
def __init__(self):
Dataset.__init__(self, data=Kemerer.data(),
dec_meta=Kemerer.decision_meta(),
obj_meta=Kemerer.objective_meta())
def __init__(self):
Dataset.__init__(self,
data=Desharnais.data(),
dec_meta=Desharnais.decision_meta(),
obj_meta=Desharnais.objective_meta())
def __init__(self):
Dataset.__init__(self, data=China.data(),
dec_meta=China.decision_meta(),
obj_meta=China.objective_meta())
def __init__(self):
Dataset.__init__(self, data=Finnish.data(),
dec_meta=Finnish.decision_meta(),
obj_meta=Finnish.objective_meta())
def __init__(self):
Dataset.__init__(self,
data=Maxwell.data(),
dec_meta=Maxwell.decision_meta(),
obj_meta=Maxwell.objective_meta())
def __init__(self, content):
Dataset.__init__(self,
limit_sentences=None,
dataset_cache_dir=None,
dataset_name=None)
self.c = content
def __init__(self):
Dataset.__init__(self,
data=Kitchenhamm.data(),
dec_meta=Kitchenhamm.decision_meta(),
obj_meta=Kitchenhamm.objective_meta())
def __init__(self):
Dataset.__init__(self,
data=Albrecht.data(),
dec_meta=Albrecht.decision_meta(),
obj_meta=Albrecht.objective_meta())
def __init__(self):
Dataset.__init__(self, data=Miyazaki.data(),
dec_meta=Miyazaki.decision_meta(),
obj_meta=Miyazaki.objective_meta())
def __init__(self, root="/ssd/data/phototourism/orig", txt_fn="data/train_phototourism_ms.txt"):
Dataset.__init__(self)
self.root = root
self.txt_fn = txt_fn
self.imgs = [line.rstrip("\n") for line in open(self.txt_fn)]
self.nimg = len(self.imgs)
