动手学深度学习PyTorch版---lastDay_4

lastDay

GAN；DCGAN

GAN相关学习连接：

GAN的变种     https://github.com/hindupuravinash/the-gan-zoo     （github,stars 7000）

“sakura小樱”博客   https://blog.csdn.net/Sakura55/article/details/81512600

知乎上  https://blog.csdn.net/Sakura55/article/details/81512600

 

1.GAN----2014

RNN,用判别模型来做生成模型

GAN，

https://www.boyuai.com/elites/course/cZu18YmweLv10OeV/jupyter/2dQ7K5yOLT4Dea6TrN6Fb

1.1 Generator

Our generator network will be the simplest network possible - a single layer linear model. This is since we will be driving that linear network with a Gaussian data generator. Hence, it literally only needs to learn the parameters to fake things perfectly.

class net_G(nn.Module):
    def __init__(self):
        super(net_G,self).__init__()
        self.model=nn.Sequential(
            nn.Linear(2,2),
        )
        self._initialize_weights()
    def forward(self,x):
        x=self.model(x)
        return x
    def _initialize_weights(self):
        for m in self.modules():
            if isinstance(m,nn.Linear):
                m.weight.data.normal_(0,0.02)
                m.bias.data.zero_()

1.2 Discriminator

For the discriminator we will be a bit more discriminating: we will use an MLP with 3 layers to make things a bit more interesting.

class net_D(nn.Module):
    def __init__(self):
        super(net_D,self).__init__()
        self.model=nn.Sequential(
            nn.Linear(2,5),
            nn.Tanh(),
            nn.Linear(5,3),
            nn.Tanh(),
            nn.Linear(3,1),
            nn.Sigmoid()
        )
        self._initialize_weights()
    def forward(self,x):
        x=self.model(x)
        return x
    def _initialize_weights(self):
        for m in self.modules():
            if isinstance(m,nn.Linear):
                m.weight.data.normal_(0,0.02)
                m.bias.data.zero_()

1.3 Training

1.3.1 定义函数更新判别器

真实数据，label=1

生成数据，label=0

# Saved in the d2l package for later use
def update_D(X,Z,net_D,net_G,loss,trainer_D):
    batch_size=X.shape[0]
    Tensor=torch.FloatTensor
    ones=Variable(Tensor(np.ones(batch_size))).view(batch_size,1)
    zeros = Variable(Tensor(np.zeros(batch_size))).view(batch_size,1)
    real_Y=net_D(X.float())
    fake_X=net_G(Z)
    fake_Y=net_D(fake_X)
    loss_D=(loss(real_Y,ones)+loss(fake_Y,zeros))/2
    loss_D.backward()
    trainer_D.step()
    return float(loss_D.sum())

1.3.2