dcgan生成一维数据代码

以下是使用PyTorch实现DCGAN生成一维数据的示例代码：

import torch
import torch.nn as nn
import torch.optim as optim
import numpy as np

# 定义生成器网络
class Generator(nn.Module):
    def __init__(self, input_size, hidden_size, output_size):
        super(Generator, self).__init__()
        self.fc1 = nn.Linear(input_size, hidden_size)
        self.relu = nn.ReLU()
        self.fc2 = nn.Linear(hidden_size, output_size)

    def forward(self, x):
        x = self.fc1(x)
        x = self.relu(x)
        x = self.fc2(x)
        return x

# 定义判别器网络
class Discriminator(nn.Module):
    def __init__(self, input_size, hidden_size, output_size):
        super(Discriminator, self).__init__()
        self.fc1 = nn.Linear(input_size, hidden_size)
        self.relu = nn.ReLU()
        self.fc2 = nn.Linear(hidden_size, output_size)
        self.sigmoid = nn.Sigmoid()

    def forward(self, x):
        x = self.fc1(x)
        x = self.relu(x)
        x = self.fc2(x)
        x = self.sigmoid(x)
        return x

# 定义训练函数
def train_dcgan(generator, discriminator, dataloader, num_epochs, lr, device):
    # 将网络移动到指定的设备上
    generator.to(device)
    discriminator.to(device)

    # 定义损失函数和优化器
    criterion = nn.BCELoss()
    g_optimizer = optim.Adam(generator.parameters(), lr=lr)
    d_optimizer = optim.Adam(discriminator.parameters(), lr=lr)

    # 定义真实标签和生成标签
    real_label = 1
    fake_label = 0

    for epoch in range(num_epochs):
        for i, data in enumerate(dataloader, 0):
            # 更新判别器网络
            discriminator.zero_grad()

            # 训练判别器网络使用真实数据
            real_data = data[0].to(device)
            batch_size = real_data.size(0)
            label = torch.full((batch_size,), real_label, device=device)
            output = discriminator(real_data)
            d_loss_real = criterion(output, label)
            d_loss_real.backward()

            # 训练判别器网络使用生成数据
            noise = torch.randn(batch_size, 100, device=device)
            fake_data = generator(noise)
            label.fill_(fake_label)
            output = discriminator(fake_data.detach())
            d_loss_fake = criterion(output, label)
            d_loss_fake.backward()

            # 更新判别器网络的参数
            d_loss = d_loss_real + d_loss_fake
            d_optimizer.step()

            # 更新生成器网络
            generator.zero_grad()
            label.fill_(real_label)
            output = discriminator(fake_data)
            g_loss = criterion(output, label)
            g_loss.backward()

            # 更新生成器网络的参数
            g_optimizer.step()

            # 打印损失值
            if i % 100 == 0:
                print('[%d/%d][%d/%d] Loss_D: %.4f Loss_G: %.4f' % (epoch, num_epochs, i, len(dataloader), d_loss.item(), g_loss.item()))

# 创建数据集
data = np.random.rand(10000, 1).astype(np.float32)
dataset = torch.utils.data.TensorDataset(torch.from_numpy(data))
dataloader = torch.utils.data.DataLoader(dataset, batch_size=64, shuffle=True)

# 定义超参数
input_size = 100
hidden_size = 64
output_size = 1
num_epochs = 50
lr = 0.0002
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

# 创建生成器和判别器网络实例
generator = Generator(input_size, hidden_size, output_size)
discriminator = Discriminator(output_size, hidden_size, 1)

# 训练DCGAN模型
train_dcgan(generator, discriminator, dataloader, num_epochs, lr, device)

在此示例代码中，我们使用了一个包含一个全连接层的生成器网络和一个包含两个全连接层的判别器网络。训练过程中，我们使用了随机噪声作为生成器网络的输入，并使用一个包含一个全连接层的判别器网络来判断生成的数据是否为真实数据。通过不断迭代，生成器网络可以学习到生成逼真的一维数据的技巧。