数据集
所有的数据集都是的子类， 它们实现了__getitem__和__len__方法。因此，它们都可以传递给.
transform = (
[(),
((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])
trainset = .CIFAR10(root='./data', train=True,
download=True, transform=transform)
trainloader = (trainset, batch_size=4,
shuffle=True, num_workers=2)
目前为止，收录的数据集包括：
__all__ = ('LSUN', 'LSUNClass',
'ImageFolder', 'DatasetFolder', 'FakeData',
'CocoCaptions', 'CocoDetection',
'CIFAR10', 'CIFAR100', 'EMNIST', 'FashionMNIST', 'QMNIST',
'MNIST', 'KMNIST', 'STL10', 'SVHN', 'PhotoTour', 'SEMEION',
'Omniglot', 'SBU', 'Flickr8k', 'Flickr30k',
'VOCSegmentation', 'VOCDetection', 'Cityscapes', 'ImageNet',
'Caltech101', 'Caltech256', 'CelebA', 'WIDERFace', 'SBDataset',
'VisionDataset', 'USPS', 'Kinetics400', "Kinetics", 'HMDB51', 'UCF101',
'Places365', 'Kitti', "INaturalist", "LFWPeople", "LFWPairs"
)
.CIFAR10模块
CIFAR-10和CIFAR-100为8000万张微小图像数据集的子集。它们是由 Alex Krizhevsky, Vinod Nair, and Geoffrey Hinton.收集的。
CIFAR-10数据集包含10类60000幅32x32彩色图像，每个类6000幅图像。训练图像50000张，测试图像10000张。该数据集被分为5个训练批和1个测试批，每个批包含10000张图像。测试批正好包含从每个类中随机选择的1000张图像。训练批以随机顺序包含剩余的图像，但有些训练批可能包含一个类的图像多于另一个类。在它们之间，训练批恰好包含来自每个类的5000张图像。
以下是数据集中的类，以及每个类的10张随机图片:
这些类是完全互斥的。汽车和卡车之间没有重叠。“汽车”包括轿车、越野车之类的东西。“卡车”只包括大卡车。这两项都不包括皮卡。
.CIFAR10源码
class CIFAR10(VisionDataset):
"""`CIFAR10 </~kriz/>`_ Dataset.
Args:
root (string): Root directory of dataset where directory
``cifar-10-batches-py`` exists or will be saved to if download is set to True.
train (bool, optional): If True, creates dataset from training set, otherwise
creates from test set.
transform (callable, optional): A function/transform that takes in an PIL image
and returns a transformed version. , ````
target_transform (callable, optional): A function/transform that takes in the
target and transforms it.
download (bool, optional): If true, downloads the dataset from the internet and
puts it in root directory. If dataset is already downloaded, it is not
downloaded again.
"""
base_folder = 'cifar-10-batches-py'
url = "/~kriz/"
filename = ""
tgz_md5 = 'c58f30108f718f92721af3b95e74349a'
train_list = [
['data_batch_1', 'c99cafc152244af753f735de768cd75f'],
['data_batch_2', 'd4bba439e000b95fd0a9bffe97cbabec'],
['data_batch_3', '54ebc095f3ab1f0389bbae665268c751'],
['data_batch_4', '634d18415352ddfa80567beed471001a'],
['data_batch_5', '482c414d41f54cd18b22e5b47cb7c3cb'],
]
test_list = [
['test_batch', '40351d587109b95175f43aff81a1287e'],
]
meta = {
'filename': '',
'key': 'label_names',
'md5': '5ff9c542aee3614f3951f8cda6e48888',
}
def __init__(
self,
root: str,
train: bool = True,
transform: Optional[Callable] = None,
target_transform: Optional[Callable] = None,
download: bool = False,
) -> None:
super(CIFAR10, self).__init__(root, transform=transform,
target_transform=target_transform)
= train # training set or test set
if download:
()
if not self._check_integrity():
raise RuntimeError('Dataset not found or corrupted.' +
' You can use download=True to download it')
if :
downloaded_list = self.train_list
else:
downloaded_list = self.test_list
: Any = []
= []
# now load the picked numpy arrays
for file_name, checksum in downloaded_list:
file_path = (, self.base_folder, file_name)
with open(file_path, 'rb') as f:
entry = (f, encoding='latin1')
(entry['data'])
if 'labels' in entry:
(entry['labels'])
else:
(entry['fine_labels'])
= ().reshape(-1, 3, 32, 32)
= ((0, 2, 3, 1)) # convert to HWC
self._load_meta()
def _load_meta(self) -> None:
path = (, self.base_folder, ['filename'])
if not check_integrity(path, ['md5']):
raise RuntimeError('Dataset metadata file not found or corrupted.' +
' You can use download=True to download it')
with open(path, 'rb') as infile:
data = (infile, encoding='latin1')
= data[['key']]
self.class_to_idx = {_class: i for i, _class in enumerate()}
def __getitem__(self, index: int) -> Tuple[Any, Any]:
"""
Args:
index (int): Index
Returns:
tuple: (image, target) where target is index of the target class.
"""
img, target = [index], [index]
# doing this so that it is consistent with all other datasets
# to return a PIL Image
img = (img)
if is not None:
img = (img)
if self.target_transform is not None:
target = self.target_transform(target)
return img, target
def __len__(self) -> int:
return len()
def _check_integrity(self) -> bool:
root =
for fentry in (self.train_list + self.test_list):
filename, md5 = fentry[0], fentry[1]
fpath = (root, self.base_folder, filename)
if not check_integrity(fpath, md5):
return False
return True
def download(self) -> None:
if self._check_integrity():
print('Files already downloaded and verified')
return
download_and_extract_archive(, , filename=, md5=self.tgz_md5)
def extra_repr(self) -> str:
return "Split: {}".format("Train" if is True else "Test")
参数说明
root (string):数据集所在目录的根目录 如果download设置为True。“cifar-10-batches-py '”存在，则将被保存至该目录
train ：如果为True，则从训练集创建数据集，否则从测试集创建。
transform::(bool，可选)一个接受PIL图像的函数/变换 并返回转换后的版本。 如”transforms“,“RandomCrop”
target_transform：(可调用，可选):一个作用于目的转换函数。
download (bool，可选):如果为true，则从internet下载数据集 ,将其放在根目录中。 如果数据集已经下载，则不会 再次下载。
trainset = torchvision.datasets.CIFAR10(root='./data', train=True,
download=True, transform=transform)
