acgan.py

"""
Example from Keras repo
"""

from __future__ import print_function, division

from keras.datasets import mnist
from keras.layers import Input, Dense, Reshape, Flatten, Dropout, multiply
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

import matplotlib.pyplot as plt

import numpy as np

class ACGAN():
    def __init__(self):
        self.img_rows = 28 
        self.img_cols = 28
        self.channels = 1
        self.num_classes = 10

        optimizer = Adam(0.0002, 0.5)
        losses = ['binary_crossentropy', 'sparse_categorical_crossentropy']

        # Build and compile the discriminator
        self.discriminator = self.build_discriminator()
        self.discriminator.compile(loss=losses, 
            optimizer=optimizer,
            metrics=['accuracy'])

        # Build and compile the generator
        self.generator = self.build_generator()
        self.generator.compile(loss=['binary_crossentropy'], 
            optimizer=optimizer)

        # The generator takes noise and the target label as input
        # and generates the corresponding digit of that label
        noise = Input(shape=(100,))
        label = Input(shape=(1,))
        img = self.generator([noise, label])

        # For the combined model we will only train the generator
        self.discriminator.trainable = False

        # The discriminator takes generated image as input and determines validity
        # and the label of that image
        valid, target_label = self.discriminator(img)

        # The combined model  (stacked generator and discriminator) takes
        # noise as input => generates images => determines validity 
        self.combined = Model([noise, label], [valid, target_label])
        self.combined.compile(loss=losses, 
            optimizer=optimizer)

    def build_generator(self):

        model = Sequential()

        model.add(Dense(128 * 7 * 7, activation="relu", input_dim=100))
        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(self.channels, kernel_size=3, padding='same'))
        model.add(Activation("tanh"))

        model.summary()

        noise = Input(shape=(100,))
        label = Input(shape=(1,), dtype='int32')

        label_embedding = Flatten()(Embedding(self.num_classes, 100)(label))

        input = multiply([noise, label_embedding])

        img = model(input)

        return Model([noise, label], img)

    def build_discriminator(self):

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

        model.add(Conv2D(16, kernel_size=3, strides=2, input_shape=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(LeakyReLU(alpha=0.2))
        model.add(Dropout(0.25))
        model.add(BatchNormalization(momentum=0.8))
        model.add(Conv2D(64, 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(128, 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=img_shape)

        features = model(img)

        validity = Dense(1, activation="sigmoid")(features)
        label = Dense(self.num_classes+1, activation="softmax")(features)

        return Model(img, [validity, label])

    def train(self, epochs, batch_size=128, save_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)

        half_batch = int(batch_size / 2)

        # 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
        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
        class_weights = [cw1, cw2]

        for epoch in range(epochs):

            # ---------------------
            #  Train Discriminator
            # ---------------------

            # Select a random half batch of images
            idx = np.random.randint(0, X_train.shape[0], half_batch)
            imgs = X_train[idx]
            
            noise = np.random.normal(0, 1, (half_batch, 100))
            # The labels of the digits that the generator tries to create and
            # image representation of
            sampled_labels = np.random.randint(0, 10, half_batch).reshape(-1, 1)

            # Generate a half batch of new images
            gen_imgs = self.generator.predict([noise, sampled_labels])

            valid = np.ones((half_batch, 1))
            fake = np.zeros((half_batch, 1))

            # Image labels. 0-9 if image is valid or 10 if it is generated (fake)
            img_labels = y_train[idx]
            fake_labels = 10 * np.ones(half_batch).reshape(-1, 1)

            # Train the discriminator
            d_loss_real = self.discriminator.train_on_batch(imgs, [valid, img_labels], class_weight=class_weights)
            d_loss_fake = self.discriminator.train_on_batch(gen_imgs, [fake, fake_labels], class_weight=class_weights)
            d_loss = 0.5 * np.add(d_loss_real, d_loss_fake)

            # ---------------------
            #  Train Generator
            # ---------------------

            noise = np.random.normal(0, 1, (batch_size, 100))

            valid = np.ones((batch_size, 1))
            # Generator wants discriminator to label the generated images as the intended
            # digits
            sampled_labels = np.random.randint(0, 10, batch_size).reshape(-1, 1)

            # Train the generator
            g_loss = self.combined.train_on_batch([noise, sampled_labels], [valid, sampled_labels], class_weight=class_weights)

            # 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[0]))

            # If at save interval => save generated image samples
            if epoch % save_interval == 0:
                self.save_model()
                self.save_imgs(epoch)

    def save_imgs(self, epoch):
        r, c = 2, 5
        noise = np.random.normal(0, 1, (r * c, 100))
        sampled_labels = np.arange(0, 10).reshape(-1, 1)

        gen_imgs = self.generator.predict([noise, sampled_labels])

        # Rescale images 0 - 1
        gen_imgs = 0.5 * gen_imgs + 0.5

        fig, axs = plt.subplots(r, c)
        fig.suptitle("ACGAN: Generated digits", fontsize=12)
        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].set_title("Digit: %d" % sampled_labels[cnt])
                axs[i,j].axis('off')
                cnt += 1
        fig.savefig("acgan/images/mnist_%d.png" % epoch)
        plt.close()

    def save_model(self):

        def save(model, model_name):
            model_path = "acgan/saved_model/%s.json" % model_name
            weights_path = "acgan/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_acgan_generator")
        save(self.discriminator, "mnist_acgan_discriminator")
        save(self.combined, "mnist_acgan_adversarial")


if __name__ == '__main__':
    acgan = ACGAN()
    acgan.train(epochs=6000, batch_size=32, save_interval=50)