网站设计有限公司怎么样,东莞市莞城建筑工程有限公司,网站建设汇报书 ppt,推动高质量发展心得体会文章目录 原理代码结果参考 原理
同时训练两个网络#xff1a;辨别器Discriminator 和 生成器Generator Generator是 造假者#xff0c;用来生成假数据。 Discriminator 是警察#xff0c;尽可能的分辨出来哪些是造假的#xff0c;哪些是真实的数据。
目的#xff1a;使… 文章目录 原理代码结果参考 原理
同时训练两个网络辨别器Discriminator 和 生成器Generator Generator是 造假者用来生成假数据。 Discriminator 是警察尽可能的分辨出来哪些是造假的哪些是真实的数据。
目的使得判别模型尽量犯错无法判断数据是来自真实数据还是生成出来的数据。
GAN的梯度下降训练过程 上图来源https://arxiv.org/abs/1406.2661
Train 辨别器: m a x max max l o g ( D ( x ) ) l o g ( 1 − D ( G ( z ) ) ) log(D(x)) log(1 - D(G(z))) log(D(x))log(1−D(G(z)))
Train 生成器: m i n min min l o g ( 1 − D ( G ( z ) ) ) log(1-D(G(z))) log(1−D(G(z)))
我们可以使用BCEloss来计算上述两个损失函数
BCEloss的表达式 m i n − [ y l n x ( 1 − y ) l n ( 1 − x ) ] min -[ylnx (1-y)ln(1-x)] min−[ylnx(1−y)ln(1−x)] 具体过程参加代码中注释
代码
import torch
import torch.nn as nn
import torch.optim as optim
import torchvision
import torchvision.datasets as datasets
from torch.utils.data import DataLoader
import torchvision.transforms as transforms
from torch.utils.tensorboard import SummaryWriter # to print to tensorboardclass Discriminator(nn.Module):def __init__(self, img_dim):super(Discriminator, self).__init__()self.disc nn.Sequential(nn.Linear(img_dim, 128),nn.LeakyReLU(0.1),nn.Linear(128, 1),nn.Sigmoid(),)def forward(self, x):return self.disc(x)class Generator(nn.Module):def __init__(self, z_dim, img_dim): # z_dim 噪声的维度super(Generator, self).__init__()self.gen nn.Sequential(nn.Linear(z_dim, 256),nn.LeakyReLU(0.1),nn.Linear(256, img_dim), # 28x28 - 784nn.Tanh(),)def forward(self, x):return self.gen(x)# Hyperparameters
device cuda if torch.cuda.is_available() else cpu
lr 3e-4 # 3e-4是Adam最好的学习率
z_dim 64 # 噪声维度
img_dim 784 # 28x28x1
batch_size 32
num_epochs 50disc Discriminator(img_dim).to(device)
gen Generator(z_dim, img_dim).to(device)fixed_noise torch.randn((batch_size, z_dim)).to(device)
transforms transforms.Compose( # MNIST标准化系数(0.1307,), (0.3081,)[transforms.ToTensor(), transforms.Normalize((0.1307, ), (0.3081,))] # 不同数据集就有不同的标准化系数
)dataset datasets.MNIST(rootdataset/, transformtransforms, downloadTrue)
loader DataLoader(dataset, batch_sizebatch_size, shuffleTrue)opt_disc optim.Adam(disc.parameters(), lrlr)
opt_gen optim.Adam(gen.parameters(), lrlr)
# BCE 损失
criterion nn.BCELoss()# 打开tensorboard在该目录下使用 tensorboard --logdirruns
writer_fake SummaryWriter(fruns/GAN_MNIST/fake)
writer_real SummaryWriter(fruns/GAN_MNIST/real)
step 0for epoch in range(num_epochs):for batch_idx, (real, _) in enumerate(loader):real real.view(-1, 784).to(device) # view相当于reshapebatch_size real.shape[0]### Train Discriminator: max log(D(real)) log(1 - D(G(z)))noise torch.randn(batch_size, z_dim).to(device)fake gen(noise) # G(z)disc_real disc(real).view(-1) # flatten# BCEloss的表达式min -[ylnx (1-y)ln(1-x)]# max log(D(real)) 相当于 min -log(D(real))# ones_like: 用1填充得到y1, 即可忽略 min -[ylnx (1-y)ln(1-x)]中的后一项# 得到 min -lnx,这里的x就是我们的real图片lossD_real criterion(disc_real, torch.ones_like(disc_real))disc_fake disc(fake).view(-1)# max log(1 - D(G(z))) 相当于 min -log(1 - D(G(z)))# zeros_like用0填充得到y0即可忽略 min -[ylnx (1-y)ln(1-x)]中的前一项# 得到 min -ln(1-x),这里的x就是我们的fake噪声lossD_fake criterion(disc_fake, torch.zeros_like(disc_fake))lossD (lossD_real lossD_fake) / 2disc.zero_grad()lossD.backward(retain_graphTrue)opt_disc.step()### Train Generator: min log(1-D(G(z))) -- max log(D(G(z))) -- min - log(D(G(z)))# 依然可使用BCEloss来做output disc(fake).view(-1)lossG criterion(output, torch.ones_like(output))gen.zero_grad()lossG.backward()opt_gen.step()if batch_idx 0:print(fEpoch [{epoch}/{num_epochs}] \ fLoss D: {lossD:.4f}, Loss G: {lossG:.4f})with torch.no_grad():fake gen(fixed_noise).reshape(-1, 1, 28, 28)data real.reshape(-1, 1, 28, 28)img_grid_fake torchvision.utils.make_grid(fake, normalizeTrue)img_grid_real torchvision.utils.make_grid(data, normalizeTrue)writer_fake.add_image(Mnist Fake Images, img_grid_fake, global_stepstep)writer_real.add_image(Mnist Real Images, img_grid_real, global_stepstep)step 1结果
训练50轮的的损失
Epoch [0/50] \ Loss D: 0.7366, Loss G: 0.7051
Epoch [1/50] \ Loss D: 0.2483, Loss G: 1.6877
Epoch [2/50] \ Loss D: 0.1049, Loss G: 2.4980
Epoch [3/50] \ Loss D: 0.1159, Loss G: 3.4923
Epoch [4/50] \ Loss D: 0.0400, Loss G: 3.8776
Epoch [5/50] \ Loss D: 0.0450, Loss G: 4.1703
...
Epoch [43/50] \ Loss D: 0.0022, Loss G: 7.7446
Epoch [44/50] \ Loss D: 0.0007, Loss G: 9.1281
Epoch [45/50] \ Loss D: 0.0138, Loss G: 6.2177
Epoch [46/50] \ Loss D: 0.0008, Loss G: 9.1188
Epoch [47/50] \ Loss D: 0.0025, Loss G: 8.9419
Epoch [48/50] \ Loss D: 0.0010, Loss G: 8.3315
Epoch [49/50] \ Loss D: 0.0007, Loss G: 7.8302使用
tensorboard --logdirruns打开tensorboard 可以看到效果并不好这是由于我们只是采用了简单的线性网络来做辨别器和生成器。后面的博文我们会使用更复杂的网络来训练GAN。
参考
[1] Building our first simple GAN [2] https://arxiv.org/abs/1406.2661
