In [1]:

    

from keras.datasets import cifar10
from keras.models import load_model
import numpy as np

from IPython import display
import matplotlib.pyplot as plt

% matplotlib inline
% config InlineBackend.figure_format = 'retina'


    

    




Using TensorFlow backend.

In [2]:

    

noise_size = 100

for image_class in range(10):
    for epoch in [500]:
        generator = load_model("networks/gen" + str(image_class) + "-" + str(epoch) + ".h5")

        noise = np.random.uniform(0, 1, size=[10, noise_size])
        generated_images = generator.predict(noise)

        print "Class {:d} generated images at epoch {:d}:".format(image_class, epoch)

        plt.figure(figsize=(10, 4))
        for i in range(generated_images.shape[0]):
            plt.subplot(2, 5, i+1)
            img = generated_images[i,:,:,:]
            plt.imshow(img)
            plt.axis('off')
        plt.tight_layout()
        plt.show()


    

    




Class 0 generated images at epoch 500:






    













    




Class 1 generated images at epoch 500:






    













    




Class 2 generated images at epoch 500:






    













    




Class 3 generated images at epoch 500:






    













    




Class 4 generated images at epoch 500:






    













    




Class 5 generated images at epoch 500:






    













    




Class 6 generated images at epoch 500:






    













    




Class 7 generated images at epoch 500:






    













    




Class 8 generated images at epoch 500:






    













    




Class 9 generated images at epoch 500:






    

In [ ]: