"""Provides a (random) weighted average between real and generated image samples"""

def _merge_function(self, inputs):

alpha = K.random_uniform((32, 1, 1, 1))

def __init__(self):

self.name = os.path.basename(__file__).split('.')[0].lower()

self.img_rows = 64

self.img_cols = 64

self.channels = 3

self.img_shape = (self.img_rows, self.img_cols, self.channels)

self.latent_dim = 100

# Following parameter and optimizer set as recommended in paper

self.n_critic = 5

optimizer = RMSprop(lr=0.00005)

# Build the generator and critic

self.generator = self.build_generator()

self.critic = self.build_critic()

# -------------------------------

# Construct Computational Graph

# for the Critic

# -------------------------------

# Freeze generator's layers while training critic

self.generator.trainable = False

# Image input (real sample)

real_img = Input(shape=self.img_shape)

# Noise input

z_disc = Input(shape=(100,))

# Generate image based of noise (fake sample)

fake_img = self.generator(z_disc)

# Discriminator determines validity of the real and fake images

fake = self.critic(fake_img)

valid = self.critic(real_img)

# Construct weighted average between real and fake images

interpolated_img = RandomWeightedAverage()([real_img, fake_img])

# Determine validity of weighted sample

validity_interpolated = self.critic(interpolated_img)

# Use Python partial to provide loss function with additional

# 'averaged_samples' argument

partial_gp_loss = partial(self.gradient_penalty_loss,

averaged_samples=interpolated_img)

partial_gp_loss.__name__ = 'gradient_penalty' # Keras requires function names

self.critic_model = Model(inputs=[real_img, z_disc],

outputs=[valid, fake, validity_interpolated])

self.critic_model.compile(loss=[self.wasserstein_loss,

self.wasserstein_loss,

partial_gp_loss],

optimizer=optimizer,

loss_weights=[1, 1, 10])

# -------------------------------

# Construct Computational Graph

# for Generator

# -------------------------------

# For the generator we freeze the critic's layers

self.critic.trainable = False

self.generator.trainable = True

# Sampled noise for input to generator

z_gen = Input(shape=(100,))

# Generate images based of noise

img = self.generator(z_gen)

# Discriminator determines validity

valid = self.critic(img)

# Defines generator model

self.generator_model = Model(z_gen, valid)

self.generator_model.compile(loss=self.wasserstein_loss, optimizer=optimizer)

filename_weight = ''

self.weight_path = 'weight/%s' % self.name

self.file_weight = os.path.join(self.weight_path, filename_weight)

self._load_weight()

def gradient_penalty_loss(self, y_true, y_pred, averaged_samples):

"""

Computes gradient penalty based on prediction and weighted real / fake samples

"""

gradients = K.gradients(y_pred, averaged_samples)[0]

# compute the euclidean norm by squaring ...

gradients_sqr = K.square(gradients)

# ... summing over the rows ...

gradients_sqr_sum = K.sum(gradients_sqr,

axis=np.arange(1, len(gradients_sqr.shape)))

# ... and sqrt

gradient_l2_norm = K.sqrt(gradients_sqr_sum)

# compute lambda * (1 - ||grad||)^2 still for each single sample

gradient_penalty = K.square(1 - gradient_l2_norm)

# return the mean as loss over all the batch samples

return K.mean(gradient_penalty)

def wasserstein_loss(self, y_true, y_pred):

return K.mean(y_true * y_pred)

def build_generator(self):

noise = Input(shape=(self.latent_dim,))

x = Dense(32 * self.img_rows * self.img_cols, activation='relu')(noise)

x = Reshape((self.img_rows, self.img_cols, 32))(x)

x = Conv2DTranspose(512, kernel_size=3, padding='same')(x)

x = BatchNormalization(momentum=0.8)(x)

x = Activation("relu")(x)

x = Conv2DTranspose(256, kernel_size=3, strides=1, padding="same")(x)

x = BatchNormalization(momentum=0.8)(x)

x = Activation("relu")(x)

x = Conv2DTranspose(128, kernel_size=3, strides=1, padding="same")(x)

x = BatchNormalization(momentum=0.8)(x)

x = Activation("relu")(x)

x = Conv2DTranspose(self.channels, kernel_size=3, padding="same")(x)

img = Activation("tanh")(x)

model = Model(noise, img)

model.summary()

return model

# model = Sequential()

# model.add(Dense(32 * 64 * 64, activation="relu", input_dim=self.latent_dim))

# model.add(Reshape((64, 64, 32)))

#

# # model.add(UpSampling2D())

# model.add(Conv2D(128, kernel_size=4, padding="same"))

# model.add(BatchNormalization(momentum=0.8))

# model.add(Activation("relu"))

#

# # model.add(UpSampling2D())

# model.add(Conv2D(64, kernel_size=4, padding="same"))

# model.add(BatchNormalization(momentum=0.8))

# model.add(Activation("relu"))

#

# model.add(Conv2D(self.channels, kernel_size=4, padding="same"))

# model.add(Activation("tanh"))

#

# model.summary()

#

#

# img = model(noise)

#

# return Model(noise, img)

def build_critic(self):

model = Sequential()

model.add(Conv2D(16, kernel_size=3, strides=2, input_shape=self.img_shape, padding="same"))

model.add(LeakyReLU(alpha=0.2))

model.add(Dropout(0.25))

model.add(Conv2D(32, kernel_size=3, strides=2, padding="same"))

model.add(ZeroPadding2D(padding=((0, 1), (0, 1))))

model.add(BatchNormalization(momentum=0.8))

model.add(LeakyReLU(alpha=0.2))

model.add(Dropout(0.25))

model.add(Conv2D(64, kernel_size=3, strides=2, padding="same"))

model.add(BatchNormalization(momentum=0.8))

model.add(LeakyReLU(alpha=0.2))

model.add(Dropout(0.25))

model.add(Conv2D(128, kernel_size=3, strides=1, padding="same"))

model.add(BatchNormalization(momentum=0.8))

model.add(LeakyReLU(alpha=0.2))

model.add(Dropout(0.25))

model.add(Flatten())

model.add(Dense(1))

model.summary()

img = Input(shape=self.img_shape)

validity = model(img)

return Model(img, validity)

def _load_weight(self):

if os.path.exists(self.file_weight) and os.path.isfile(self.file_weight):

self.generator_model.load_weights(self.file_weight, by_name=True)

print('Load weight [%s] successfully.' % self.file_weight)

else:

print('No weight loaded.')

def _save_weight(self, epoch, g_loss):

filename_weight = '%s_%05d-%.4f.h5' % (self.name, epoch, g_loss)

file_weight = os.path.join(self.weight_path, filename_weight)

if not os.path.exists(self.weight_path):

os.makedirs(self.weight_path)

self.generator_model.save_weights(file_weight, overwrite=True)

def train(self, epochs, batch_size, sample_interval=50):

# Load the dataset

# (X_train, _), (_, _) = mnist.load_data()

# generator = load_data(batch_size, self.img_rows, self.img_cols)

generator = load_data(batch_size, self.img_rows, self.img_cols, imgaug=True, path=ANIME_PATH)

# Rescale -1 to 1

# X_train = (X_train.astype(np.float32) - 127.5) / 127.5

# X_train = np.expand_dims(X_train, axis=3)

# Adversarial ground truths

valid = -np.ones((batch_size, 1))

fake = np.ones((batch_size, 1))

dummy = np.zeros((batch_size, 1)) # Dummy gt for gradient penalty

for epoch in range(epochs):

for _ in range(self.n_critic):

# ---------------------

# Train Discriminator

# ---------------------

# Select a random batch of images

# idx = np.random.randint(0, X_train.shape[0], batch_size)

# imgs = X_train[idx]

imgs, _ = next(generator)

imgs = imgs / 127.5 - 1.

if imgs.shape[0] != batch_size:

continue

# Sample generator input

noise = np.random.normal(0, 1, (batch_size, self.latent_dim))

# Train the critic

d_loss = self.critic_model.train_on_batch([imgs, noise],

[valid, fake, dummy])

# ---------------------

# Train Generator

# ---------------------

g_loss = self.generator_model.train_on_batch(noise, valid)

# Plot the progress

print("%d [D loss: %f] [G loss: %f]" % (epoch, d_loss[0], g_loss))

# If at save interval => save generated image samples

if epoch % sample_interval == 0:

self.sample_images(epoch)

self._save_weight(epoch, g_loss)

def sample_images(self, epoch):

r, c = 5, 5

noise = np.random.normal(0, 1, (r * c, self.latent_dim))

gen_imgs = self.generator.predict(noise)

# Rescale images 0 - 1

gen_imgs = 0.5 * gen_imgs + 1

fig, axs = plt.subplots(r, c)

cnt = 0

for i in range(r):

for j in range(c):

axs[i, j].imshow(gen_imgs[cnt, :, :, :])

axs[i, j].axis('off')

cnt += 1

save_dir = 'images/%s' % self.name

if not os.path.exists(save_dir):

os.makedirs(save_dir)

fig.savefig("%s/%d.png" % (save_dir, epoch))