无中生有之Tensorflow2.0 GAN mnist数据集 生成图像

推荐一个可爱的博主 ，专门介绍各种GAN, 还有代码讲解 https://www.bilibili.com/video/BV13J41187Fo?p=4

然后他的博客  https://blog.csdn.net/weixin_44791964/category_9625179.html 视频的代码会放在这个里面

讲一下自己的理解吧。

1.先定义生成器和鉴别器

2.然后将生成器和鉴别器连起来

3. 在训练阶段，先训练鉴别器，

4. 再训练生成器，训练生成器时，鉴别器设置为不训练self.discriminator.trainable = False

5. 训练生成器和训练鉴别器的噪声数据是不同的，都是随机生成的，生成器训练时只是用了鉴别器当前训练好的参数。

注：直接copy上面博客的代码导入两个库时会报错，要修改一下

from __future__ import print_function, division
import tensorflow as tf
from  tensorflow.keras.datasets import mnist
from  tensorflow.keras.layers import Input, Dense, Reshape, Flatten, Dropout,Conv2D
from  tensorflow.keras.layers import BatchNormalization, Activation, ZeroPadding2D, LeakyReLU
#from  tensorflow.keras.layers.advanced_activations import LeakyReLU
#from  tensorflow.keras.layers.convolutional import UpSampling2D, Conv2D
from  tensorflow.keras.models import Sequential, Model
from  tensorflow.keras.optimizers import Adam

import matplotlib.pyplot as plt

import sys
import os
import numpy as np

class GAN():
    def __init__(self):
        # --------------------------------- #
        #   行28，列28，也就是mnist的shape
        # --------------------------------- #
        self.img_rows = 28
        self.img_cols = 28
        self.channels = 1
        # 28,28,1
        self.img_shape = (self.img_rows, self.img_cols, self.channels)
        self.latent_dim = 100  #噪声的维度
        # adam优化器
        optimizer = Adam(0.0002, 0.5)
        #--------------------------训练鉴别器--------------------------
        self.discriminator = self.build_discriminator()
        self.discriminator.compile(loss='binary_crossentropy',
            optimizer=optimizer,
            metrics=['accuracy'])
        #----------------------------训练生成器-------------------------------
        self.generator = self.build_generator()
        gan_input = Input(shape=(self.latent_dim,))
        img = self.generator(gan_input)   #Model(noise, img)
        # 在训练generate的时候不训练discriminator
        self.discriminator.trainable = False
        # 对生成的假图片进行预测
        validity = self.discriminator(img) #Model(img, validity)
        self.combined = Model(gan_input, validity)  #把生成器和鉴别器连起来
        self.combined.compile(loss='binary_crossentropy', optimizer=optimizer)


    def build_generator(self):
        # --------------------------------- #
        #   生成器，输入一串随机数字，输出一副图像
        # --------------------------------- #
        model = Sequential()

        model.add(Dense(256, input_dim=self.latent_dim))
        model.add(LeakyReLU(alpha=0.2))
        model.add(BatchNormalization(momentum=0.8))

        model.add(Dense(512))
        model.add(LeakyReLU(alpha=0.2))
        model.add(BatchNormalization(momentum=0.8))

        model.add(Dense(1024))
        model.add(LeakyReLU(alpha=0.2))
        model.add(BatchNormalization(momentum=0.8))
        
        #生成图像大小的序列 28*28 *1
        model.add(Dense(np.prod(self.img_shape), activation='tanh'))
        #然后reshape成（28,28,1）大小
        model.add(Reshape(self.img_shape))

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

        return Model(noise, img)

    def build_discriminator(self):
        # ----------------------------------- #
        #   评价器，对输入进来的图片进行鉴别真假
        # ----------------------------------- #
        model = Sequential()
        # 输入一张图片
        model.add(Flatten(input_shape=self.img_shape))
        model.add(Dense(512))
        model.add(LeakyReLU(alpha=0.2))
        model.add(Dense(256))
        model.add(LeakyReLU(alpha=0.2))
        # 判断真伪，输出概率，1为真，0为假
        model.add(Dense(1, activation='sigmoid'))   

        img = Input(shape=self.img_shape)
        validity = model(img)

        return Model(img, validity)

    #训练步骤
    def train(self, epochs, batch_size=128, sample_interval=50):
        # 获得数据
        (X_train, _), (_, _) = mnist.load_data()

        # 进行标准化(-1,1)之间
        X_train = X_train / 127.5 - 1.
        X_train = np.expand_dims(X_train, axis=3)# (28,28)--->(28,28,1)

        # 创建标签
        valid = np.ones((batch_size, 1))   
        fake = np.zeros((batch_size, 1))

        for epoch in range(epochs):

            # --------------------------- #
            #   随机选取batch_size个图片
            #   对discriminator进行训练
            # --------------------------- #
            idx = np.random.randint(0, X_train.shape[0], batch_size)  #在0到总图像数量中随机产生一个batch_size的序列
            imgs = X_train[idx]     #随机选取图像

            noise = np.random.normal(0, 1, (batch_size, self.latent_dim)) #在（0,1）中随机产生batch_size个latent_dim维的噪声

            gen_imgs = self.generator.predict(noise)  #由噪声生成图片
            #计算鉴别器loss = 0.5 *（鉴别真图像的loss +  鉴别假图像的loss）
            d_loss_real = self.discriminator.train_on_batch(imgs, valid)
            d_loss_fake = self.discriminator.train_on_batch(gen_imgs, fake)
            d_loss = 0.5 * np.add(d_loss_real, d_loss_fake)  

            # --------------------------- #
            #  训练generator
            # --------------------------- #
            noise = np.random.normal(0, 1, (batch_size, self.latent_dim))   #为了增加generator的随机性，这里的噪声是重新生成的，只是用了训练好的参数
            g_loss = self.combined.train_on_batch(noise, valid)
            print ("%d [D loss: %f, acc.: %.2f%%] [G loss: %f]" % (epoch, d_loss[0], 100*d_loss[1], g_loss))

            if epoch % sample_interval == 0:
                self.sample_images(epoch)

    def sample_images(self, epoch):
        #采样25个图像
        r, c = 5, 5
        noise = np.random.normal(0, 1, (r * c, self.latent_dim))
        gen_imgs = self.generator.predict(noise) #为了增加generator的随机性，这里的噪声是重新生成的，只是用了训练好的参数

        gen_imgs = 0.5 * gen_imgs + 0.5   #生成的图像在（-1,1）之间，转化到（0,1）之间

        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/%d.png" % epoch)
        plt.close()


if __name__ == '__main__':
    if not os.path.exists("./images"):
        os.makedirs("./images")
    gan = GAN()
    gan.train(epochs=30000, batch_size=256, sample_interval=200)

看我无中生有的图片：第一张是0epoch的，第二张是两万多epoch的

直接在笔记本上面跑的，还挺快的，也就只能跑这种比较简单的数据集了，等我有钱了，我一定要买最好的设备，哼