Semi-supervised generative adversarial networks (SGAN)

from __future__ import print_function, division
 
from keras.datasets import mnist
from keras.layers import Input, Dense, Reshape, Flatten, Dropout, multiply, GaussianNoise
from keras.layers import BatchNormalization, Activation, Embedding, ZeroPadding2D
from keras.layers.advanced_activations import LeakyReLU
from keras.layers.convolutional import UpSampling2D, Conv2D
from keras.models import Sequential, Model
from keras.optimizers import Adam
from keras import losses
from keras.utils import to_categorical
import keras.backend as K
 
import matplotlib.pyplot as plt
 
import numpy as np
 
class SGAN:
    def __init__(self):
        self.img_rows = 28
        self.img_cols = 28
        self.channels = 1
        self.img_shape = (self.img_rows, self.img_cols, self.channels)
        self.num_classes = 10
        self.latent_dim = 100
 
        optimizer = Adam(0.0002, 0.5)
 
        # Build and compile the discriminator
        self.discriminator = self.build_discriminator()
        self.discriminator.compile(
            loss=['binary_crossentropy', 'categorical_crossentropy'],
            loss_weights=[0.5, 0.5],
            optimizer=optimizer,
            metrics=['accuracy']
        )
 
        # Build the generator
        self.generator = self.build_generator()
 
        # The generator takes noise as input and generates imgs
        noise = Input(shape=(100,))
        img = self.generator(noise)
 
        # For the combined model we will only train the generator
        self.discriminator.trainable = False
 
        # The valid takes generated images as input and determines validity
        valid, _ = self.discriminator(img)
 
        # The combined model  (stacked generator and discriminator)
        # Trains generator to fool discriminator
        self.combined = Model(noise, valid)
        self.combined.compile(loss=['binary_crossentropy'], optimizer=optimizer)
 
    def build_generator(self):
 
        model = Sequential()
 
        model.add(Dense(128 * 7 * 7, activation="relu", input_dim=self.latent_dim))
        model.add(Reshape((7, 7, 128)))
        model.add(BatchNormalization(momentum=0.8))
        model.add(UpSampling2D())
        model.add(Conv2D(128, kernel_size=3, padding="same"))
        model.add(Activation("relu"))
        model.add(BatchNormalization(momentum=0.8))
        model.add(UpSampling2D())
        model.add(Conv2D(64, kernel_size=3, padding="same"))
        model.add(Activation("relu"))
        model.add(BatchNormalization(momentum=0.8))
        model.add(Conv2D(1, kernel_size=3, padding="same"))
        model.add(Activation("tanh"))
 
        model.summary()
 
        noise = Input(shape=(self.latent_dim,))
        img = model(noise)
 
        return Model(noise, img)
 
    def build_discriminator(self):
 
        model = Sequential()
 
        model.add(Conv2D(32, 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(64, kernel_size=3, strides=2, padding="same"))
        model.add(ZeroPadding2D(padding=((0,1),(0,1))))
        model.add(LeakyReLU(alpha=0.2))
        model.add(Dropout(0.25))
        model.add(BatchNormalization(momentum=0.8))
        model.add(Conv2D(128, kernel_size=3, strides=2, padding="same"))
        model.add(LeakyReLU(alpha=0.2))
        model.add(Dropout(0.25))
        model.add(BatchNormalization(momentum=0.8))
        model.add(Conv2D(256, kernel_size=3, strides=1, padding="same"))
        model.add(LeakyReLU(alpha=0.2))
        model.add(Dropout(0.25))
        model.add(Flatten())
 
        model.summary()
 
        img = Input(shape=self.img_shape)
 
        features = model(img)
        valid = Dense(1, activation="sigmoid")(features)
        label = Dense(self.num_classes+1, activation="softmax")(features)
 
        return Model(img, [valid, label])
 
    def train(self, epochs, batch_size=128, sample_interval=50):
 
        # Load the dataset
        (X_train, y_train), (_, _) = mnist.load_data()
 
        # Rescale -1 to 1
        X_train = (X_train.astype(np.float32) - 127.5) / 127.5
        X_train = np.expand_dims(X_train, axis=3)
        y_train = y_train.reshape(-1, 1)
 
        # Class weights:
        # To balance the difference in occurences of digit class labels.
        # 50% of labels that the discriminator trains on are 'fake'.
        # Weight = 1 / frequency
        half_batch = batch_size // 2
        cw1 = {0: 1, 1: 1}
        cw2 = {i: self.num_classes / half_batch for i in range(self.num_classes)}
        cw2[self.num_classes] = 1 / half_batch
 
        # Adversarial ground truths
        valid = np.ones((batch_size, 1))
        fake = np.zeros((batch_size, 1))
 
        for epoch in range(epochs):
 
            # ---------------------
            #  Train Discriminator
            # ---------------------
 
            # Select a random batch of images
            idx = np.random.randint(0, X_train.shape[0], batch_size)
            imgs = X_train[idx]
 
            # Sample noise and generate a batch of new images
            noise = np.random.normal(0, 1, (batch_size, self.latent_dim))
            gen_imgs = self.generator.predict(noise)
 
            # One-hot encoding of labels
            labels = to_categorical(y_train[idx], num_classes=self.num_classes+1)
            fake_labels = to_categorical(np.full((batch_size, 1), self.num_classes), num_classes=self.num_classes+1)
 
            # Train the discriminator
            d_loss_real = self.discriminator.train_on_batch(imgs, [valid, labels], class_weight=[cw1, cw2])
            d_loss_fake = self.discriminator.train_on_batch(gen_imgs, [fake, fake_labels], class_weight=[cw1, cw2])
            d_loss = 0.5 * np.add(d_loss_real, d_loss_fake)
 
 
            # ---------------------
            #  Train Generator
            # ---------------------
 
            g_loss = self.combined.train_on_batch(noise, valid, class_weight=[cw1, cw2])
 
            # Plot the progress
            print ("%d [D loss: %f, acc: %.2f%%, op_acc: %.2f%%] [G loss: %f]" % (epoch, d_loss[0], 100*d_loss[3], 100*d_loss[4], g_loss))
 
            # If at save interval => save generated image samples
            if epoch % sample_interval == 0:
                self.sample_images(epoch)
 
    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 + 0.5
 
        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, :,:,0], cmap='gray')
                axs[i,j].axis('off')
                cnt += 1
        fig.savefig("images/mnist_%d.png" % epoch)
        plt.close()
 
    def save_model(self):
 
        def save(model, model_name):
            model_path = "saved_model/%s.json" % model_name
            weights_path = "saved_model/%s_weights.hdf5" % model_name
            options = {"file_arch": model_path,
                        "file_weight": weights_path}
            json_string = model.to_json()
            open(options['file_arch'], 'w').write(json_string)
            model.save_weights(options['file_weight'])
 
        save(self.generator, "mnist_sgan_generator")
        save(self.discriminator, "mnist_sgan_discriminator")
        save(self.combined, "mnist_sgan_adversarial")
 
 
if __name__ == '__main__':
    sgan = SGAN()
    sgan.train(epochs=20000, batch_size=32, sample_interval=50)