cf0181bb0bf5add0686ca4dd4c03e6fb04a34703,snntoolbox/io_utils/datasets/cifar10.py,,get_cifar10,#Any#Any#Any#,21

Before Change

//       np.savez_compressed(filepath+"Y_train", Y_train)
        np.savez_compressed(filepath+"Y_test", Y_test)

    return (X_train, Y_train, X_test, Y_test)

After Change

    

    // Whether to apply global contrast normalization and ZCA whitening
    gcn = True
    zca = True
    nb_classes = 10

    (X_train, y_train), (X_test, y_test) = cifar10.load_data()

    // Convert class vectors to binary class matrices
    Y_train = to_categorical(y_train, nb_classes)
    Y_test = to_categorical(y_test, nb_classes)

    datagen = ImageDataGenerator(rescale=1./255, featurewise_center=gcn,
                                 featurewise_std_normalization=gcn,
                                 zca_whitening=zca)
    datagen.fit(X_test/255.)
    dataflow = datagen.flow(X_test, Y_test, batch_size=len(X_test))
    X_test, Y_test = dataflow.next()

    if flat:
        X_train = X_train.reshape(X_train.shape[0], np.prod(X_train.shape[1:]))
        X_test = X_test.reshape(X_test.shape[0], np.prod(X_test.shape[1:]))

Instances