diff --git a/MNIST_GAN/MNIST_GAN.py b/MNIST_GAN/MNIST_GAN.py index f7451fa57594cc277d486c7c84f4bbf612f4a481..cc0d7e983e0cca5f7fde4b08ffa71cd4c97dbac0 100644 --- a/MNIST_GAN/MNIST_GAN.py +++ b/MNIST_GAN/MNIST_GAN.py @@ -72,14 +72,14 @@ y[ntrain:, 0] = 1 # class 0 for generated images discriminator.fit(X, y, epochs=1, batch_size=128) # - Create a dataset of 5000 real test images and 5000 fake images. -no = np.random.choice(10000, size=ntrain/2, replace='False') +no = np.random.choice(10000, size=ntrain//2, replace='False') real_test = X_test[no, :, :, :] # sample real images from test set -noise_gen = np.random.uniform(0, 1, size=[ntrain/2, latent_dim]) +noise_gen = np.random.uniform(0, 1, size=[ntrain//2, latent_dim]) generated_images = generator.predict(noise_gen) # generate fake images with untrained generator Xt = np.concatenate((real_test, generated_images)) yt = np.zeros([ntrain, 2]) # class vector: one-hot encoding -yt[:ntrain/2, 1] = 1 # class 1 for real images -yt[ntrain/2:, 0] = 1 # class 0 for generated images +yt[:ntrain//2, 1] = 1 # class 1 for real images +yt[ntrain//2:, 0] = 1 # class 0 for generated images # - Evaluate the test accuracy of your network. pretrain_loss, pretrain_acc = discriminator.evaluate(Xt, yt, verbose=0, batch_size=128) @@ -99,7 +99,7 @@ def train_for_n(epochs=1, batch_size=32): generated_images = generator.predict(noise) plot_images(generated_images, fname=log_dir + '/generated_images_' + str(epoch)) - iterations_per_epoch = 60000/batch_size # number of training steps per epoch + iterations_per_epoch = 60000//batch_size # number of training steps per epoch perm = np.random.choice(60000, size=60000, replace='False') for i in range(iterations_per_epoch):