生成对抗网络——GAN深度卷积实现（代码+理解）_卷积实现生成器和判别器

作者：繁依Fanyi0 | 2024-07-24 18:14:27

踩

卷积实现生成器和判别器

本篇博客为上篇博客的另一个实现版本，训练流程相同，所以只实现代码，感兴趣可以跳转看一下。

生成对抗网络—GAN（代码+理解）

http://t.csdnimg.cn/HDfLOhttp://t.csdnimg.cn/HDfLO

一、GAN深度卷积实现

1. 模型结构

（1）生成器（Generator）

（2）判别器（Discriminator）

一、GAN深度卷积实现

1. 模型结构

（1）生成器（Generator）

（2）判别器（Discriminator）

2. 代码实现


import torch
import torch.nn as nn
import argparse
import torchvision.transforms as transforms
import matplotlib.pyplot as plt
from torch.utils.data import DataLoader
from torchvision import datasets
import numpy as np
 
 
parser = argparse.ArgumentParser()
parser.add_argument("--n_epochs", type=int, default=20, help="number of epochs of training")
parser.add_argument("--batch_size", type=int, default=64, help="size of the batches")
parser.add_argument("--lr", type=float, default=0.0002, help="adam: learning rate")
parser.add_argument("--b1", type=float, default=0.5, help="adam: decay of first order momentum of gradient")
parser.add_argument("--b2", type=float, default=0.999, help="adam: decay of first order momentum of gradient")
parser.add_argument("--latent_dim", type=int, default=100, help="dimensionality of the latent space")
parser.add_argument("--img_size", type=int, default=32, help="size of each image dimension")
parser.add_argument("--channels", type=int, default=1, help="number of image channels")
parser.add_argument("--sample_interval", type=int, default=400, help="interval between image sampling")
opt = parser.parse_args()
print(opt)
 
# 加载数据
dataloader = torch.utils.data.DataLoader(
    datasets.MNIST(
        "./others/",
        train=False,
        download=False,
        transform=transforms.Compose(
            [transforms.Resize(opt.img_size), transforms.ToTensor(), transforms.Normalize([0.5], [0.5])]
        ),
    ),
    batch_size=opt.batch_size,
    shuffle=True,
)
 
def weights_init_normal(m):
    classname = m.__class__.__name__
    if classname.find("Conv") != -1:
        torch.nn.init.normal_(m.weight.data, 0.0, 0.02)
    elif classname.find("BatchNorm2d") != -1:
        torch.nn.init.normal_(m.weight.data, 1.0, 0.02) # 给定均值和标准差的正态分布N(mean,std)中生成值
        torch.nn.init.constant_(m.bias.data, 0.0)
 
class Generator(nn.Module):
    def __init__(self):
        super(Generator, self).__init__()
 
        self.init_size = opt.img_size // 4  # 原为28*28，现为32*32，两边各多了2
        self.l1 = nn.Sequential(nn.Linear(opt.latent_dim, 128 * self.init_size ** 2))
 
        self.conv_blocks = nn.Sequential(
            nn.BatchNorm2d(128),    # 调整数据的分布，使其 更适合于 下一层的 激活函数或学习
            nn.Upsample(scale_factor=2),
            nn.Conv2d(128, 128, 3, stride=1, padding=1),
            nn.BatchNorm2d(128, 0.8),
            nn.LeakyReLU(0.2, inplace=True),
            nn.Upsample(scale_factor=2),
            nn.Conv2d(128, 64, 3, stride=1, padding=1),
            nn.BatchNorm2d(64, 0.8),
            nn.LeakyReLU(0.2, inplace=True),
            nn.Conv2d(64, opt.channels, 3, stride=1, padding=1),
            nn.Tanh(),
        )
 
    def forward(self, z):
        out = self.l1(z)
        out = out.view(out.shape[0], 128, self.init_size, self.init_size)
        img = self.conv_blocks(out)
        return img
 
class Discriminator(nn.Module):
    def __init__(self):
        super(Discriminator, self).__init__()
 
        def discriminator_block(in_filters, out_filters, bn=True):
            block = [nn.Conv2d(in_filters, out_filters, 3, 2, 1),
                     nn.LeakyReLU(0.2, inplace=True),
                     nn.Dropout2d(0.25)]
            if bn:
                block.append(nn.BatchNorm2d(out_filters, 0.8))
            return block
 
        self.model = nn.Sequential(
            *discriminator_block(opt.channels, 16, bn=False),
            *discriminator_block(16, 32),
            *discriminator_block(32, 64),
            *discriminator_block(64, 128),
        )
        # 下采样（图片进行 4次卷积操作，变为ds_size * ds_size尺寸大小）
        ds_size = opt.img_size // 2 ** 4
        self.adv_layer = nn.Sequential(
            nn.Linear(128 * ds_size ** 2, 1),
            nn.Sigmoid()
        )
 
    def forward(self, img):
        out = self.model(img)
        out = out.view(out.shape[0], -1)
        validity = self.adv_layer(out)
        return validity
 
# 实例化
generator = Generator()
discriminator = Discriminator()
 
# 初始化参数
generator.apply(weights_init_normal)
discriminator.apply(weights_init_normal)
 
# 优化器
optimizer_G = torch.optim.Adam(generator.parameters(), lr=opt.lr, betas=(opt.b1, opt.b2))
optimizer_D = torch.optim.Adam(discriminator.parameters(), lr=opt.lr, betas=(opt.b1, opt.b2))
 
# 交叉熵损失函数
adversarial_loss = torch.nn.BCELoss()
 
def gen_img_plot(model, epoch, text_input):
    prediction = np.squeeze(model(text_input).detach().cpu().numpy()[:16])
    plt.figure(figsize=(4, 4))
    for i in range(16):
        plt.subplot(4, 4, i + 1)
        plt.imshow((prediction[i] + 1) / 2)
        plt.axis('off')
    plt.show()
 
# ----------
#  Training
# ----------
D_loss_ = []  # 记录训练过程中判别器的损失
G_loss_ = []  # 记录训练过程中生成器的损失
for epoch in range(opt.n_epochs):
    # 初始化损失值
    D_epoch_loss = 0
    G_epoch_loss = 0
    count = len(dataloader)  # 返回批次数
    for i, (imgs, _) in enumerate(dataloader):
        valid = torch.ones(imgs.shape[0], 1)
        fake = torch.zeros(imgs.shape[0], 1)
 
        # -----------------
        #  Train Generator
        # -----------------
        optimizer_G.zero_grad()
        z = torch.randn(imgs.shape[0], opt.latent_dim)
        gen_imgs = generator(z)
        g_loss = adversarial_loss(discriminator(gen_imgs), valid)
        g_loss.backward()
        optimizer_G.step()
 
        # ---------------------
        #  Train Discriminator
        # ---------------------
        optimizer_D.zero_grad()
        real_loss = adversarial_loss(discriminator(imgs), valid)
        fake_loss = adversarial_loss(discriminator(gen_imgs.detach()), fake)
        d_loss = (real_loss + fake_loss) / 2
        d_loss.backward()
        optimizer_D.step()
 
        print(
            "[Epoch %d/%d] [Batch %d/%d] [D loss: %f] [G loss: %f]"
            % (epoch, opt.n_epochs, i, len(dataloader), d_loss.item(), g_loss.item())
        )
 
        # batches_done = epoch * len(dataloader) + i
        # if batches_done % opt.sample_interval == 0:
        #     save_image(gen_imgs.data[:25], "others/images/%d.png" % batches_done, nrow=5, normalize=True)
 
        # 累计每一个批次的loss
        with torch.no_grad():
            D_epoch_loss += d_loss
            G_epoch_loss += g_loss
 
        # 求平均损失
    with torch.no_grad():
        D_epoch_loss /= count
        G_epoch_loss /= count
        D_loss_.append(D_epoch_loss.item())
        G_loss_.append(G_epoch_loss.item())
 
        text_input = torch.randn(opt.batch_size, opt.latent_dim)
        gen_img_plot(generator, epoch, text_input)
 
 
x = [epoch + 1 for epoch in range(opt.n_epochs)]
plt.figure()
plt.plot(x, G_loss_, 'r')
plt.plot(x, D_loss_, 'b')
plt.ylabel('loss')
plt.xlabel('epoch')
plt.legend(['G_loss','D_loss'])
plt.show()