3 DCGAN

Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks

3.1 简介

深度卷积神经网络（Deep Convolution Generative Adversarial Nets）的生成网络

在DCGAN中，

判别网络是一个传统的深度卷积神经网络，但使用了带步长的卷积来实现下采样操作，不用 m a x p o o l i n g maxpooling maxpooling操作；
生成网络使用了一个特殊的深度卷积网络来实现。如上图，使用微步卷积来生成 64 × 64 64×64 64×64大小的图像。第一层是全连接层，输入是从均匀分布中随机采样的100维向量 z z z，输出是 4 × 4 × 1024 4×4×1024 4×4×1024的向量，后面是四层微步卷积，没有汇聚层。

DCGAN的主要优点是通过一些经验性的网络结构设计使得对抗训练更加稳定。比如：

使用带步长的卷积（在判别网络中）和微步卷积（在生成网络中）来代替汇聚操作，以免损失信息；
使用批量归一化；
去除卷积层之后的全连接层；
在生成网络中，除了最后一层使用了 T a n h Tanh Tanh激活函数以外,其余层都使用 R e L U ReLU ReLU函数；
在判别网络中，都是用 L e a k y R e L U LeakyReLU LeakyReLU函数。

3.2 DGGAN实现

# DCGAN_2016.py
import os
import numpy as np
import matplotlib.pyplot as plt
from tensorflow.keras.models import Sequential, Model
from tensorflow.keras.layers import Dense, LeakyReLU, BatchNormalization, Reshape, Input
from tensorflow.keras.layers import UpSampling2D, Conv2D, Activation, ZeroPadding2D, GlobalAveragePooling2D
from tensorflow.keras.datasets import mnist
from tensorflow.keras.optimizers import Adamclass DCGAN():def __init__(self):# 输入shapeself.img_rows = 28self.img_cols = 28self.channels = 1self.img_shape = (self.img_rows, self.img_cols, self.channels)# 分十类self.num_classes = 10self.latent_dim = 100# adam优化器optimizer = Adam(0.0002, 0.5)# 判别模型self.discriminator = self.build_discriminator()self.discriminator.compile(loss=['binary_crossentropy'],optimizer=optimizer,metrics=['accuracy'])# 生成模型self.generator = self.build_generator()# conbine是生成模型和判别模型的结合# 判别模型的trainable为False# 用于训练生成模型z = Input(shape=(self.latent_dim,))img = self.generator(z)self.discriminator.trainable = Falsevalid = self.discriminator(img)self.combined = Model(z, valid)self.combined.compile(loss='binary_crossentropy', optimizer=optimizer)def build_generator(self):model = Sequential()# 先全连接到64*7*7的维度上model.add(Dense(32 * 7 * 7, activation="relu", input_dim=self.latent_dim))# reshape成特征层的样式model.add(Reshape((7, 7, 32)))# 7, 7, 64model.add(Conv2D(64, kernel_size=3, padding="same"))model.add(BatchNormalization(momentum=0.8))model.add(Activation("relu"))# 上采样# 7, 7, 64 -> 14, 14, 64model.add(UpSampling2D())model.add(Conv2D(128, kernel_size=3, padding="same"))model.add(BatchNormalization(momentum=0.8))model.add(Activation("relu"))# 上采样# 14, 14, 128 -> 28, 28, 64model.add(UpSampling2D())model.add(Conv2D(64, kernel_size=3, padding="same"))model.add(BatchNormalization(momentum=0.8))model.add(Activation("relu"))# 上采样# 28, 28, 64 -> 28, 28, 1model.add(Conv2D(self.channels, 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()# 28, 28, 1 -> 14, 14, 32model.add(Conv2D(32, kernel_size=3, strides=2, input_shape=self.img_shape, padding="same"))model.add(BatchNormalization(momentum=0.8))model.add(LeakyReLU(alpha=0.2))# 14, 14, 32 -> 7, 7, 64model.add(Conv2D(64, kernel_size=3, strides=2, padding="same"))model.add(BatchNormalization(momentum=0.8))model.add(LeakyReLU(alpha=0.2))# 7, 7, 64 -> 4, 4, 128model.add(ZeroPadding2D(((0, 1), (0, 1))))model.add(Conv2D(128, kernel_size=3, strides=2, padding="same"))model.add(BatchNormalization(momentum=0.8))model.add(LeakyReLU(alpha=0.2))model.add(GlobalAveragePooling2D())# 全连接model.add(Dense(1, activation='sigmoid'))model.summary()img = Input(shape=self.img_shape)validity = model(img)return Model(img, validity)def train(self, epochs, batch_size=128, save_interval=50):(X_train, _), (_, _) = mnist.load_data()       # 载入数据X_train = X_train / 127.5 - 1.  # 归一化X_train = np.expand_dims(X_train, axis=3)# Adversarial ground truthsvalid = np.ones((batch_size, 1))fake = np.zeros((batch_size, 1))for epoch in range(epochs):# --------------------- ##  训练判别模型# --------------------- #idx = np.random.randint(0, X_train.shape[0], batch_size)imgs = X_train[idx]noise = np.random.normal(0, 1, (batch_size, self.latent_dim))gen_imgs = self.generator.predict(noise)# 训练并计算lossd_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)# ---------------------#  训练生成模型# ---------------------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 % save_interval == 0:self.save_imgs(epoch)def save_imgs(self, epoch):r, c = 5, 5noise = np.random.normal(0, 1, (r * c, self.latent_dim))gen_imgs = self.generator.predict(noise)gen_imgs = 0.5 * gen_imgs + 0.5fig, axs = plt.subplots(r, c)cnt = 0for 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 += 1fig.savefig("images/mnist_%d.png" % epoch)plt.close()if __name__ == '__main__':# if not os.path.exists("./images"):#     os.makedirs("./images")dcgan = DCGAN()dcgan.train(epochs=20000, batch_size=256, save_interval=500)

【GANs】Deep Convolution Generative Adversarial Network相关推荐

【GAN】《ENERGY-BASED GENERATIVE ADVERSARIAL NETWORKS》 ICLR‘17
<ENERGY-BASED GENERATIVE ADVERSARIAL NETWORKS> 先介绍EBGAN,再详细解读为什么这样做. Auto-encoder Discriminato ...
Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network论文翻译——中英文对照
文章作者:Tyan 博客:noahsnail.com | CSDN | 简书声明:作者翻译论文仅为学习,如有侵权请联系作者删除博文,谢谢! 翻译论文汇总:https://github.com ...
【论文翻译】中英对照翻译--（Attentive Generative Adversarial Network for Raindrop Removal from A Single Image）
[开始时间]2018.10.08 [完成时间]2018.10.09 [论文翻译]Attentive GAN论文中英对照翻译--(Attentive Generative Adversarial Net ...
【GANs】Generative Adversarial Nets
[GANs]Generative Adversarial Nets 1 GAN 1.1 GANs的简介 1.2 思想与目标函数 1.3 GAN代码 1.4 全局最优推导 1.5 GANs方向展望 1 ...
【GANs】Conditional Generative Adversarial Nets
[GANs]Conditional Generative Adversarial Nets 2 CGAN 2.1 CGAN简介前言流程图目标函数 2.2 CGAN代码 2 CGAN 2.1 CG ...
【论文笔记】DOA-GAN: Dual-Order Attentive Generative Adversarial Network for Image Copy-move Forgery Detec
DOA-GAN: Dual-Order Attentive Generative Adversarial Network for Image Copy-move Forgery Detection a ...
【深度学习】Generative Adversarial Network 生成式对抗网络（GAN）
文章目录一.神经网络作为生成器 1.1 什么是生成器? 1.2 为什么需要输出一个分布? 1.3 什么时候需要生成器? 二.Generative Adversarial Network 生成式对抗网 ...
GANs学习系列(8)：Deep Convolutional Generative Adversarial Nerworks，DCGAN
[前言] 本文首先介绍生成式模型,然后着重梳理生成式模型(Generative Models)中生成对抗网络(Generative Adversarial Network)的研究与发展.作者 ...
Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network
Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network Ledig C, Theis ...

【GANs】Deep Convolution Generative Adversarial Network

【GANs】Deep Convolution Generative Adversarial Network

3 DCGAN

3.1 简介

3.2 DGGAN实现

【GANs】Deep Convolution Generative Adversarial Network相关推荐

最新文章

热门文章