Cifar100

class imagepypelines.core.ml.Cifar100(fraction=1)[source]

Bases: imagepypelines.core.ml.datasets.DatasetRetrieval

Object to load the cifar100 dataset in a pipeline compatible format

Parameters:label_mode (string) – ‘fine’ for individual labels (100 unique), ‘coarse’ for superclass labels (20 unique)
x_train

50,000 color 32,32,3 images

Type:list
y_train

corresponding integer labels for the data

Type:list
x_test

10,000 color 32,32,3 images

Type:list
y_test

corresponding integer labels for the data

Type:list

Methods Summary

get_sorted_test() retrieves data and labels for test set sorted by label
get_sorted_train() retrieves data and labels for train set sorted by label
get_test() retrieves the cifar 100 labels test dataset using keras
get_train() retrieves the cifar 100 labels train dataset using keras

Methods Documentation

get_sorted_test()

retrieves data and labels for test set sorted by label

get_sorted_train()

retrieves data and labels for train set sorted by label

get_test()[source]

retrieves the cifar 100 labels test dataset using keras

Parameters:None
Returns:10,000 color 32,32,3 images y_test(list): corresponding integer labels for the data
Return type:x_test(list)
get_train()[source]

retrieves the cifar 100 labels train dataset using keras

Parameters:None
Returns:50,000 color 32,32,3 images y_train(list): corresponding integer labels for the data
Return type:x_train(list)