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: cifar100
• keywords: image_processing, classification
• dataset size: 355,3 MB
• is downloadable: yes
• tasks: classification (default)

Tasks

classification (default)

• How to use
• Properties
• HDF5 file structure
• Fields

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 classification
• description: Contains image tensors and label annotations for image classification.
• sets: train, test
• metadata file size in disk: 177,8 MB

• has annotations: yes

  • which:

    • labels for each image class/category.

• available fields:

  • classes
  • superclasses
  • images
  • labels
  • coarse_labels
  • object_fields
  • object_ids
  • list_images_per_class
  • list_images_per_superclass

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 descriptions

  • available in: train, test
  • dtype: np.uint8
  • is padded: True
  • fill value: 0
  • note: strings stored in ASCII format

• superclasses: super class names. It is composed of groups of classes per super class

  • available in: train, test
  • dtype: np.uint8
  • is padded: True
  • fill value: 0
  • note: strings stored in ASCII format

• images: images tensor

  • available in: train, test
  • dtype: np.uint8
  • is padded: False
  • fill value: -1

• labels: class ids

  • available in: train, test
  • dtype: np.uint8
  • is padded: False
  • fill value: -1

• coarse_labels: superclass ids

  • available in: train, test
  • dtype: np.uint8
  • is padded: False
  • fill value: -1

• object_fields: list of field names of the object id list

  • available in: train, test
  • dtype: np.uint8
  • is padded: True
  • fill value: 0
  • note: strings stored in ASCII format
  • note: key field (field name aggregator)

• object_ids: list of field ids

  • available in: train, test
  • dtype: np.int32
  • is padded: False
  • fill value: -1
  • note: key field (field id aggregator)

• list_images_per_class: list of image ids per class

  • available in: train, test
  • dtype: np.int32
  • is padded: True
  • fill value: -1
  • note: pre-ordered list

• list_images_per_superclass: list of image ids per superclass

  • available in: train, test
  • dtype: np.int32
  • is padded: True
  • fill value: -1
  • note: pre-ordered list

Disclaimer

All rights reserved to the original creators of CIFAR-100.

For information about the dataset and its terms of use, please see this link.