CIFAR-100¶
This dataset is just like the CIFAR-10, except it has 100 classes containing 600 images each. There are 500 training images and 100 testing images per class. The 100 classes in the CIFAR-100 are grouped into 20 superclasses. Each image comes with a “fine” label (the class to which it belongs) and a “coarse” label (the superclass to which it belongs).
Use cases¶
Image classification.
Properties¶
name
: cifar100keywords
: image_processing, classificationdataset size
: 355,3 MBis downloadable
: yestasks
: classification (default)
Tasks¶
classification (default)¶
How to use¶
>>> # import the package
>>> import dbcollection as dbc
>>>
>>> # load the dataset
>>> cifar100 = dbc.load('cifar100')
>>> cifar100
DataLoader: "cifar100" (classification task)
Properties¶
primary use
: image classificationdescription
: Contains image tensors and label annotations for image classification.sets
: train, testmetadata file size in disk
: 177,8 MBhas annotations
: yeswhich
:- labels for each image class/category.
HDF5 file structure¶
/
├── train/
│ ├── classes # dtype=np.uint8, shape=(100,18) (note: string in ASCII format)
│ ├── superclasses # dtype=np.uint8, shape=(20,31) (note: string in ASCII format)
│ ├── images # dtype=np.uint8, shape=(50000,32,32,3)
│ ├── labels # dtype=np.uint8, shape=(50000,)
│ ├── coarse_labels # dtype=np.uint8, shape=(50000,)
│ ├── object_fields # dtype=np.uint8, shape=(3,13) (note: string in ASCII format)
│ ├── object_ids # dtype=np.int32, shape=(50000,3)
│ ├── list_images_per_class # dtype=np.int32, shape=(100,500))
│ └── list_images_per_superclass # dtype=np.int32, shape=(20,2500))
│
└── test/
├── classes # dtype=np.uint8, shape=(100,18) (note: string in ASCII format)
├── superclasses # dtype=np.uint8, shape=(20,31) (note: string in ASCII format)
├── images # dtype=np.uint8, shape=(10000,32,32,3)
├── labels # dtype=np.uint8, shape=(10000,)
├── coarse_labels # dtype=np.uint8, shape=(10000,)
├── object_fields # dtype=np.uint8, shape=(3,13) (note: string in ASCII format)
├── object_ids # dtype=np.int32, shape=(10000,3)
├── list_images_per_class # dtype=np.int32, shape=(100,100))
└── list_images_per_superclass # dtype=np.int32, shape=(20,500))
Fields¶
classes
: class descriptionsavailable in
: train, testdtype
: np.uint8is padded
: Truefill value
: 0note
: strings stored in ASCII format
superclasses
: super class names. It is composed of groups of classes per super classavailable in
: train, testdtype
: np.uint8is padded
: Truefill value
: 0note
: strings stored in ASCII format
images
: images tensoravailable in
: train, testdtype
: np.uint8is padded
: Falsefill value
: -1
labels
: class idsavailable in
: train, testdtype
: np.uint8is padded
: Falsefill value
: -1
coarse_labels
: superclass idsavailable in
: train, testdtype
: np.uint8is padded
: Falsefill value
: -1
object_fields
: list of field names of the object id listavailable in
: train, testdtype
: np.uint8is padded
: Truefill value
: 0note
: strings stored in ASCII formatnote
: key field (field name aggregator)
object_ids
: list of field idsavailable in
: train, testdtype
: np.int32is padded
: Falsefill value
: -1note
: key field (field id aggregator)
list_images_per_class
: list of image ids per classavailable in
: train, testdtype
: np.int32is padded
: Truefill value
: -1note
: pre-ordered list
list_images_per_superclass
: list of image ids per superclassavailable in
: train, testdtype
: np.int32is padded
: Truefill value
: -1note
: pre-ordered list