[深度学习-实践]GAN入门例子-利用Tensorflow Keras与数据集CIFAR10生成新图片

系列文章目录

深度学习GAN（一）之简单介绍
深度学习GAN（二）之基于CIFAR10数据集的例子;
深度学习GAN（三）之基于手写体Mnist数据集的例子;
深度学习GAN（四）之PIX2PIX GAN的例子

---

0. 前言

CIFAR-10 数据集介绍请看这篇博客
GPU怎么配置请看这篇博客

通过本博客你知道怎么利用GAN产生大量的假的图片、
如下图，左边是真实的图片。右边是利用GAN产生的图片

下图是GAN生成的手写体数字，用了10个epoch

1. GAN的简单介绍

生成式对抗网络（GAN, Generative Adversarial Networks ）是一种深度学习模型，是近年来复杂分布上无监督学习最具前景的方法之一。模型通过框架中（至少）两个模块：生成模型（Generative Model）和判别模型（Discriminative Model）的互相博弈学习产生相当好的输出。原始 GAN 理论中，并不要求 G 和 D 都是神经网络，只需要是能拟合相应生成和判别的函数即可。但实用中一般均使用深度神经网络作为 G 和 D 。一个优秀的GAN应用需要有良好的训练方法，否则可能由于神经网络模型的自由性而导致输出不理想。

要了解GAN的工作原理以及如何在GAN架构中训练深层卷积神经网络模型以生成图像，可能是具有挑战性的。对于初学者来说，一个很好的起点是在计算机视觉领域中使用的标准图像数据集（例如CIFAR小物体照片数据集）上练习开发和使用GAN。使用小型且易于理解的数据集意味着可以快速开发和训练较小的模型，从而可以将重点放在模型体系结构和图像生成过程本身上。

在本博客中，您将发现如何使用具有深度卷积网络的生成对抗网络来生成物体的小照片。

2. GAN原理

在训练过程中，生成器努力地让生成的图像更加真实，而判别器则努力地去识别出图像的真假，这个过程相当与一个二人博弈，随着时间的推移，生成器和判别器在不断地进行对抗。

下图是GAN的整个结构的图片

3. 如何定义与训练判别器（discriminator model）

3.1 定义判别器模型

定义一个判别器模型的函数define_discriminator()，输入参数是图片的大小，默认（32，32，3）

import tensorflow as tf
import tensorflow.keras as keras
import numpy as np
import matplotlib.pyplot as plt

# define the standalone discriminator model
def define_discriminator(in_shape=(32,32,3)):
    model = keras.models.Sequential()
    # normal
    model.add(keras.layers.Conv2D(64, (3,3), padding='same', input_shape=in_shape))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    # downsample
    model.add(keras.layers.Conv2D(128, (3,3), strides=(2,2), padding='same'))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    # downsample
    model.add(keras.layers.Conv2D(128, (3,3), strides=(2,2), padding='same'))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    # downsample
    model.add(keras.layers.Conv2D(256, (3,3), strides=(2,2), padding='same'))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    # classifier
    model.add(keras.layers.Flatten())
    model.add(keras.layers.Dropout(0.4))
    model.add(keras.layers.Dense(1, activation='sigmoid'))
    # compile model
    opt = keras.optimizers.Adam(lr=0.0002, beta_1=0.5)
    model.compile(loss='binary_crossentropy', optimizer=opt, metrics=['accuracy'])

    return model
if __name__ == '__main__':  
	model = define_discriminator()
	# summarize the model
	model.summary()

结果

Model: "sequential"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
conv2d (Conv2D)              (None, 32, 32, 64)        1792      
_________________________________________________________________
leaky_re_lu (LeakyReLU)      (None, 32, 32, 64)        0         
_________________________________________________________________
conv2d_1 (Conv2D)            (None, 16, 16, 128)       73856     
_________________________________________________________________
leaky_re_lu_1 (LeakyReLU)    (None, 16, 16, 128)       0         
_________________________________________________________________
conv2d_2 (Conv2D)            (None, 8, 8, 128)         147584    
_________________________________________________________________
leaky_re_lu_2 (LeakyReLU)    (None, 8, 8, 128)         0         
_________________________________________________________________
conv2d_3 (Conv2D)            (None, 4, 4, 256)         295168    
_________________________________________________________________
leaky_re_lu_3 (LeakyReLU)    (None, 4, 4, 256)         0         
_________________________________________________________________
flatten (Flatten)            (None, 4096)              0         
_________________________________________________________________
dropout (Dropout)            (None, 4096)              0         
_________________________________________________________________
dense (Dense)                (None, 1)                 4097      
=================================================================
Total params: 522,497
Trainable params: 522,497
Non-trainable params: 0

3.2. 数据集归一化已经转化为浮点类型

把图片的像素点从[0,255]转成[-1, 1]

# load and prepare cifar10 training images
def load_real_samples():
    # load cifar10 dataset
    (trainX, _), (_, _) = tf.keras.datasets.cifar10.load_data()
    # convert from unsigned ints to floats
    X = trainX.astype('float32')
    # scale from [0,255] to [-1,1]
    X = (X - 127.5) / 127.5
    return X

3.3. 随机的取n_sample个真实图片从数据集中

Y 都是1

# select real samples
def generate_real_samples(dataset, n_samples):
    # choose random instances
    ix = np.random.randint(0, dataset.shape[0], n_samples)
    # retrieve selected images
    X = dataset[ix]
    # generate 'real' class labels (1)
    y = np.ones((n_samples, 1))
    return X, y

3.4. 随机制造n_sample个假的图片

Y都是0

def generate_fake_samples1(n_samples):
    # generate uniform random numbers in [0,1]
    X = np.random.rand(32 * 32 * 3 * n_samples)
    # update to have the range [-1, 1]
    X = -1 + X * 2
    # reshape into a batch of color images
    X = X.reshape((n_samples, 32, 32, 3))
    # generate 'fake' class labels (0)
    y = np.zeros((n_samples, 1))
    return X, y

3.5.训练判别器

真实图片与假的图片一起训练。

# train the discriminator model
def train_discriminator(model, dataset, n_iter=20, n_batch=128):
    half_batch = int(n_batch / 2)
    # manually enumerate epochs
    for i in range(n_iter):
        # get randomly selected 'real' samples
        X_real, y_real = generate_real_samples(dataset, half_batch)
        # update discriminator on real samples
        _, real_acc = model.train_on_batch(X_real, y_real)
        # generate 'fake' examples
        X_fake, y_fake = generate_fake_samples1(half_batch)
        # update discriminator on fake samples
        _, fake_acc = model.train_on_batch(X_fake, y_fake)
        # summarize performance
        print('>%d real=%.0f%% fake=%.0f%%' % (i+1, real_acc*100, fake_acc*100))
def test_train_discriminator():
    # define the discriminator model
    model = define_discriminator()
    # load image data
    dataset = load_real_samples()
    # fit the model
    train_discriminator(model, dataset)
if __name__ == '__main__':
    test_train_discriminator()

结果如下，可以看出训练的时候真实的图片的准确率与假的图片的准确率

>1 real=83% fake=0%
>2 real=100% fake=0%
>3 real=100% fake=8%
>4 real=94% fake=31%
>5 real=97% fake=62%
>6 real=86% fake=92%
>7 real=98% fake=98%
>8 real=92% fake=100%
>9 real=88% fake=100%
>10 real=94% fake=100%
>11 real=94% fake=100%
>12 real=100% fake=100%
>13 real=91% fake=100%
>14 real=97% fake=100%
>15 real=98% fake=100%
>16 real=100% fake=100%
>17 real=97% fake=100%
>18 real=100% fake=100%
>19 real=100% fake=100%
>20 real=100% fake=100%

4. 如何定义与使用产生器(Generator)

4.1 定义产生器

产生器的生成正好和判别器是相反的。
是由100维的向量反向生成（32，32，3）的图片
此处的Generator不用指定loss函数与优化器，因为它不能直接训练，
下面还有其它的。

# define the standalone generator model
def define_generator(latent_dim):
    model = keras.models.Sequential()
    # foundation for 4x4 image
    n_nodes = 256 * 4 * 4
    model.add(keras.layers.Dense(n_nodes, input_dim=latent_dim))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    model.add(keras.layers.Reshape((4, 4, 256)))
    # upsample to 8x8
    model.add(keras.layers.Conv2DTranspose(128, (4,4), strides=(2,2), padding='same'))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    # upsample to 16x16
    model.add(keras.layers.Conv2DTranspose(128, (4,4), strides=(2,2), padding='same'))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    # upsample to 32x32
    model.add(keras.layers.Conv2DTranspose(128, (4,4), strides=(2,2), padding='same'))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    # output layer
    model.add(keras.layers.Conv2D(3, (3,3), activation='tanh', padding='same'))
    return model
# define the size of the latent space
latent_dim = 100
# define the generator model
model = define_generator(latent_dim)
# summarize the model
model.summary()

执行结果如下，

Model: "sequential"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
dense (Dense)                (None, 4096)              413696    
_________________________________________________________________
leaky_re_lu (LeakyReLU)      (None, 4096)              0         
_________________________________________________________________
reshape (Reshape)            (None, 4, 4, 256)         0         
_________________________________________________________________
conv2d_transpose (Conv2DTran (None, 8, 8, 128)         524416    
_________________________________________________________________
leaky_re_lu_1 (LeakyReLU)    (None, 8, 8, 128)         0         
_________________________________________________________________
conv2d_transpose_1 (Conv2DTr (None, 16, 16, 128)       262272    
_________________________________________________________________
leaky_re_lu_2 (LeakyReLU)    (None, 16, 16, 128)       0         
_________________________________________________________________
conv2d_transpose_2 (Conv2DTr (None, 32, 32, 128)       262272    
_________________________________________________________________
leaky_re_lu_3 (LeakyReLU)    (None, 32, 32, 128)       0         
_________________________________________________________________
conv2d (Conv2D)              (None, 32, 32, 3)         3459      
=================================================================
Total params: 1,466,115
Trainable params: 1,466,115
Non-trainable params: 0

4.2 展示假的图片

generate_latent_points（）产生n_samples个随机向量，每个向量默认为100维
generate_fake_samples（）根据产生器生成假的图片
show_fake_sample（）展示假的图片

# generate points in latent space as input for the generator
def generate_latent_points(latent_dim, n_samples):
    # generate points in the latent space
    x_input = np.random.randn(latent_dim * n_samples)
    # reshape into a batch of inputs for the network
    x_input = x_input.reshape(n_samples, latent_dim)
    return x_input


# use the generator to generate n fake examples, with class labels
def generate_fake_samples(g_model, latent_dim, n_samples):
    # generate points in latent space
    x_input = generate_latent_points(latent_dim, n_samples)
    # predict outputs
    X = g_model.predict(x_input)
    # create 'fake' class labels (0)
    y = np.zeros((n_samples, 1))
    return X, y


def show_fake_sample():
    # size of the latent space
    latent_dim = 100
    # define the discriminator model
    model = define_generator(latent_dim)
    # generate samples
    n_samples = 49
    X, _ = generate_fake_samples(model, latent_dim, n_samples)
    # scale pixel values from [-1,1] to [0,1]
    X = (X + 1) / 2.0
    # plot the generated samples
    for i in range(n_samples):
        # define subplot
        plt.subplot(7, 7, 1 + i)
        # turn off axis labels
        plt.axis('off')
        # plot single image
        plt.imshow(X[i])
    # show the figure
    plt.show()
if __name__ == '__main__':
    show_fake_sample()

执行结果如下，因为还没有训练，所以图片都是灰色的。

4.3 定义GAN模型

Gan的模型其实是由Generator与Desciminator组成的

# define the combined generator and discriminator model, for updating the generator
def define_gan(g_model, d_model):
    # make weights in the discriminator not trainable
    d_model.trainable = False
    # connect them
    model = tf.keras.models.Sequential()
    # add generator
    model.add(g_model)
    # add the discriminator
    model.add(d_model)
    # compile model
    opt = tf.keras.optimizers.Adam(lr=0.0002, beta_1=0.5)
    model.compile(loss='binary_crossentropy', optimizer=opt)
    return model
def show_gan_module():
    # size of the latent space
    latent_dim = 100
    # create the discriminator
    d_model = define_discriminator()
    # create the generator
    g_model = define_generator(latent_dim)
    # create the gan
    gan_model = define_gan(g_model, d_model)
    # summarize gan model
    gan_model.summary()
if __name__ == '__main__':
    show_gan_module()

执行结果如下：

Model: "sequential_2"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
sequential_1 (Sequential)    (None, 32, 32, 3)         1466115   
_________________________________________________________________
sequential (Sequential)      (None, 1)                 522497    
=================================================================
Total params: 1,988,612
Trainable params: 1,466,115
Non-trainable params: 522,497

4.4. 训练GAN模型

# evaluate the discriminator, plot generated images, save generator model
def summarize_performance(epoch, g_model, d_model, dataset, latent_dim, n_samples=150):
    # prepare real samples
    X_real, y_real = generate_real_samples(dataset, n_samples)
    # evaluate discriminator on real examples
    _, acc_real = d_model.evaluate(X_real, y_real, verbose=0)
    # prepare fake examples
    x_fake, y_fake = generate_fake_samples(g_model, latent_dim, n_samples)
    # evaluate discriminator on fake examples
    _, acc_fake = d_model.evaluate(x_fake, y_fake, verbose=0)
    # summarize discriminator performance
    print('>Accuracy real: %.0f%%, fake: %.0f%%' % (acc_real * 100, acc_fake * 100))
    # save plot
    #save_plot(x_fake, epoch)
    # save the generator model tile file
    filename = 'generator_model_%03d.h5' % (epoch + 1)
    g_model.save(filename)


# train the generator and discriminator
def train(g_model, d_model, gan_model, dataset, latent_dim, n_epochs=200, n_batch=128):
    bat_per_epo = int(dataset.shape[0] / n_batch)
    half_batch = int(n_batch / 2)
    # manually enumerate epochs
    for i in range(n_epochs):
        # enumerate batches over the training set
        for j in range(bat_per_epo):
            # get randomly selected 'real' samples
            X_real, y_real = generate_real_samples(dataset, half_batch)
            # update discriminator model weights
            d_loss1, _ = d_model.train_on_batch(X_real, y_real)
            # generate 'fake' examples
            X_fake, y_fake = generate_fake_samples(g_model, latent_dim, half_batch)
            # update discriminator model weights
            d_loss2, _ = d_model.train_on_batch(X_fake, y_fake)
            # prepare points in latent space as input for the generator
            X_gan = generate_latent_points(latent_dim, n_batch)
            # create inverted labels for the fake samples
            y_gan = np.ones((n_batch, 1))
            # update the generator via the discriminator's error
            g_loss = gan_model.train_on_batch(X_gan, y_gan)
            # summarize loss on this batch
            print('>%d, %d/%d, d1=%.3f, d2=%.3f g=%.3f' %
                  (i + 1, j + 1, bat_per_epo, d_loss1, d_loss2, g_loss))
        # evaluate the model performance, sometimes
        if (i + 1) % 10 == 0:
            summarize_performance(i, g_model, d_model, dataset, latent_dim)


def train_gan():
    # size of the latent space
    latent_dim = 100
    # create the discriminator
    d_model = define_discriminator()
    # create the generator
    g_model = define_generator(latent_dim)
    # create the gan
    gan_model = define_gan(g_model, d_model)
    # load image data
    dataset = load_real_samples()
    # train model
    train(g_model, d_model, gan_model, dataset, latent_dim)
if __name__ == '__main__':
    train_gan()

5.展示Generator产生的假的图片

加载训练好的权重参数，随机产生100个100维的点。利用Generator产生假图片并展示出来
注意：模型的权重参数文件名字有可能要改变一下。

def show_imgs_for_final_generator_model():
    # load model
    model = tf.keras.models.load_model('generator_model_010.h5')
    # generate images
    latent_points = generate_latent_points(100, 100)
    # generate images
    X = model.predict(latent_points)
    # scale from [-1,1] to [0,1]
    X = (X + 1) / 2.0
    # plot the result
    create_plot(X, 10)

10个epoch之后，结果如下：感觉虽然每个图片有都有不同。但是还是很难分别它是什么。

90，100个epoch之后。我们就可以识别出具体的图片是什么了。

100 epoch之后， 就稳定了

6 显示一张假图片

利用一个所有值都是0.75的向量产生一个图片

def show_single_imgs():
    model = tf.keras.models.load_model('generator_model_200.h5')
    # all 0s
    vector = np.asarray([[0.75 for _ in range(100)]])
    # generate image
    X = model.predict(vector)
    # scale from [-1,1] to [0,1]
    X = (X + 1) / 2.0
    # plot the result
    plt.imshow(X[0, :, :])
    plt.show()

结果如下，看起来像鹿，又像是鹿与马组成的动物。

7. GAN基于CIFAR10数据集的完整代码

import tensorflow as tf
import tensorflow.keras as keras
import numpy as np
import matplotlib.pyplot as plt

# define the standalone discriminator model
def define_discriminator(in_shape=(32,32,3)):
    model = keras.models.Sequential()
    # normal
    model.add(keras.layers.Conv2D(64, (3,3), padding='same', input_shape=in_shape))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    # downsample
    model.add(keras.layers.Conv2D(128, (3,3), strides=(2,2), padding='same'))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    # downsample
    model.add(keras.layers.Conv2D(128, (3,3), strides=(2,2), padding='same'))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    # downsample
    model.add(keras.layers.Conv2D(256, (3,3), strides=(2,2), padding='same'))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    # classifier
    model.add(keras.layers.Flatten())
    model.add(keras.layers.Dropout(0.4))
    model.add(keras.layers.Dense(1, activation='sigmoid'))
    # compile model
    opt = keras.optimizers.Adam(lr=0.0002, beta_1=0.5)
    model.compile(loss='binary_crossentropy', optimizer=opt, metrics=['accuracy'])

    model.summary()
    return model


# load and prepare cifar10 training images
def load_real_samples():
    # load cifar10 dataset
    (trainX, _), (_, _) = tf.keras.datasets.cifar10.load_data()
    # convert from unsigned ints to floats
    X = trainX.astype('float32')
    # scale from [0,255] to [-1,1]
    X = (X - 127.5) / 127.5
    return X


# select real samples
def generate_real_samples(dataset, n_samples):
    # choose random instances
    ix = np.random.randint(0, dataset.shape[0], n_samples)
    # retrieve selected images
    X = dataset[ix]
    # generate 'real' class labels (1)
    y = np.ones((n_samples, 1))
    return X, y


def generate_fake_samples1(n_samples):
    # generate uniform random numbers in [0,1]
    X = np.random.rand(32 * 32 * 3 * n_samples)
    # update to have the range [-1, 1]
    X = -1 + X * 2
    # reshape into a batch of color images
    X = X.reshape((n_samples, 32, 32, 3))
    # generate 'fake' class labels (0)
    y = np.zeros((n_samples, 1))
    return X, y


# train the discriminator model
def train_discriminator(model, dataset, n_iter=20, n_batch=128):
    half_batch = int(n_batch / 2)
    # manually enumerate epochs
    for i in range(n_iter):
        # get randomly selected 'real' samples
        X_real, y_real = generate_real_samples(dataset, half_batch)
        # update discriminator on real samples
        _, real_acc = model.train_on_batch(X_real, y_real)
        # generate 'fake' examples
        X_fake, y_fake = generate_fake_samples1(half_batch)
        # update discriminator on fake samples
        _, fake_acc = model.train_on_batch(X_fake, y_fake)
        # summarize performance
        print('>%d real=%.0f%% fake=%.0f%%' % (i+1, real_acc*100, fake_acc*100))

def test_train_discriminator():
    # define the discriminator model
    model = define_discriminator()
    # load image data
    dataset = load_real_samples()
    # fit the model
    train_discriminator(model, dataset)


# define the standalone generator model
def define_generator(latent_dim):
    model = keras.models.Sequential()
    # foundation for 4x4 image
    n_nodes = 256 * 4 * 4
    model.add(keras.layers.Dense(n_nodes, input_dim=latent_dim))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    model.add(keras.layers.Reshape((4, 4, 256)))
    # upsample to 8x8
    model.add(keras.layers.Conv2DTranspose(128, (4,4), strides=(2,2), padding='same'))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    # upsample to 16x16
    model.add(keras.layers.Conv2DTranspose(128, (4,4), strides=(2,2), padding='same'))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    # upsample to 32x32
    model.add(keras.layers.Conv2DTranspose(128, (4,4), strides=(2,2), padding='same'))
    model.add(keras.layers.LeakyReLU(alpha=0.2))
    # output layer
    model.add(keras.layers.Conv2D(3, (3,3), activation='tanh', padding='same'))
    return model

# generate points in latent space as input for the generator
def generate_latent_points(latent_dim, n_samples):
    # generate points in the latent space
    x_input = np.random.randn(latent_dim * n_samples)
    # reshape into a batch of inputs for the network
    x_input = x_input.reshape(n_samples, latent_dim)
    return x_input


# use the generator to generate n fake examples, with class labels
def generate_fake_samples(g_model, latent_dim, n_samples):
    # generate points in latent space
    x_input = generate_latent_points(latent_dim, n_samples)
    # predict outputs
    X = g_model.predict(x_input)
    # create 'fake' class labels (0)
    y = np.zeros((n_samples, 1))
    return X, y


def show_fake_sample():
    # size of the latent space
    latent_dim = 100
    # define the discriminator model
    model = define_generator(latent_dim)
    # generate samples
    n_samples = 49
    X, _ = generate_fake_samples(model, latent_dim, n_samples)
    # scale pixel values from [-1,1] to [0,1]
    X = (X + 1) / 2.0
    # plot the generated samples
    for i in range(n_samples):
        # define subplot
        plt.subplot(7, 7, 1 + i)
        # turn off axis labels
        plt.axis('off')
        # plot single image
        plt.imshow(X[i])
    # show the figure
    plt.show()


# define the combined generator and discriminator model, for updating the generator
def define_gan(g_model, d_model):
    # make weights in the discriminator not trainable
    d_model.trainable = False
    # connect them
    model = tf.keras.models.Sequential()
    # add generator
    model.add(g_model)
    # add the discriminator
    model.add(d_model)
    # compile model
    opt = tf.keras.optimizers.Adam(lr=0.0002, beta_1=0.5)
    model.compile(loss='binary_crossentropy', optimizer=opt)
    return model

def show_gan_module():
    # size of the latent space
    latent_dim = 100
    # create the discriminator
    d_model = define_discriminator()
    # create the generator
    g_model = define_generator(latent_dim)
    # create the gan
    gan_model = define_gan(g_model, d_model)
    # summarize gan model
    gan_model.summary()



# evaluate the discriminator, plot generated images, save generator model
def summarize_performance(epoch, g_model, d_model, dataset, latent_dim, n_samples=150):
    # prepare real samples
    X_real, y_real = generate_real_samples(dataset, n_samples)
    # evaluate discriminator on real examples
    _, acc_real = d_model.evaluate(X_real, y_real, verbose=0)
    # prepare fake examples
    x_fake, y_fake = generate_fake_samples(g_model, latent_dim, n_samples)
    # evaluate discriminator on fake examples
    _, acc_fake = d_model.evaluate(x_fake, y_fake, verbose=0)
    # summarize discriminator performance
    print('>Accuracy real: %.0f%%, fake: %.0f%%' % (acc_real * 100, acc_fake * 100))
    # save plot
    #save_plot(x_fake, epoch)
    # save the generator model tile file
    filename = 'generator_model_%03d.h5' % (epoch + 1)
    g_model.save(filename)


# train the generator and discriminator
def train(g_model, d_model, gan_model, dataset, latent_dim, n_epochs=200, n_batch=128):
    bat_per_epo = int(dataset.shape[0] / n_batch)
    half_batch = int(n_batch / 2)
    # manually enumerate epochs
    for i in range(n_epochs):
        # enumerate batches over the training set
        for j in range(bat_per_epo):
            # get randomly selected 'real' samples
            X_real, y_real = generate_real_samples(dataset, half_batch)
            # update discriminator model weights
            d_loss1, _ = d_model.train_on_batch(X_real, y_real)
            # generate 'fake' examples
            X_fake, y_fake = generate_fake_samples(g_model, latent_dim, half_batch)
            # update discriminator model weights
            d_loss2, _ = d_model.train_on_batch(X_fake, y_fake)
            # prepare points in latent space as input for the generator
            X_gan = generate_latent_points(latent_dim, n_batch)
            # create inverted labels for the fake samples
            y_gan = np.ones((n_batch, 1))
            # update the generator via the discriminator's error
            g_loss = gan_model.train_on_batch(X_gan, y_gan)
            # summarize loss on this batch
            print('>%d, %d/%d, d1=%.3f, d2=%.3f g=%.3f' %
                  (i + 1, j + 1, bat_per_epo, d_loss1, d_loss2, g_loss))
        # evaluate the model performance, sometimes
        if (i + 1) % 10 == 0:
            summarize_performance(i, g_model, d_model, dataset, latent_dim)


def train_gan():
    # size of the latent space
    latent_dim = 100
    # create the discriminator
    d_model = define_discriminator()
    # create the generator
    g_model = define_generator(latent_dim)
    # create the gan
    gan_model = define_gan(g_model, d_model)
    # load image data
    dataset = load_real_samples()
    # train model
    train(g_model, d_model, gan_model, dataset, latent_dim)



# generate points in latent space as input for the generator
def generate_latent_points(latent_dim, n_samples):
    # generate points in the latent space
    x_input = np.random.randn(latent_dim * n_samples)
    # reshape into a batch of inputs for the network
    x_input = x_input.reshape(n_samples, latent_dim)
    return x_input

# plot the generated images
def create_plot(examples, n):
    # plot images
    for i in range(n * n):
        # define subplot
        plt.subplot(n, n, 1 + i)
        # turn off axis
        plt.axis('off')
        # plot raw pixel data
        plt.imshow(examples[i, :, :])
    plt.show()

def show_imgs_for_final_generator_model():
    # load model
    model = tf.keras.models.load_model('generator_model_010.h5')
    # generate images
    latent_points = generate_latent_points(100, 100)
    # generate images
    X = model.predict(latent_points)
    # scale from [-1,1] to [0,1]
    X = (X + 1) / 2.0
    # plot the result
    create_plot(X, 10)

def show_single_imgs():
    model = tf.keras.models.load_model('generator_model_200.h5')
    # all 0s
    vector = np.asarray([[0.75 for _ in range(100)]])
    # generate image
    X = model.predict(vector)
    # scale from [-1,1] to [0,1]
    X = (X + 1) / 2.0
    # plot the result
    plt.imshow(X[0, :, :])
    plt.show()

if __name__ == '__main__':
    #define_discriminator()
    #test_train_discriminator()
   # show_fake_sample()
    #show_gan_module()
    train_gan()
    show_imgs_for_final_generator_model()