有关条件GAN（cgan)的相关原理，可以参考：
GAN系列之CGAN原理简介以及pytorch项目代码实现
其他类型的GAN原理介绍以及应用，可以查看我的GANs专栏
一、数据集介绍，加载数据
依旧使用到的是我们的老朋友-----MNIST手写数字数据集, 本文不再详细做介绍
相关数据集介绍可以参考:深度学习入门--MNIST数据集及创建自己的手写数字数据集
传统GAN生成手写数字参考：入门GAN实战---生成MNIST手写数据集代码实现pytorch
DCGAN生成手写数字参考：Pytorch 使用DCGAN生成MNIST手写数字 入门级教程
# 独热编码def one_hot(x, class_count=10):return torch.eye(class_count)[x, :] transform =pose([transforms.ToTensor(),transforms.Normalize(0.5, 0.5)])# 这个数据集其实包含两部分，第一部分是数据，第二部分是标签 print（dataset[0])#这个第二部分就是我们所需要的condition，这个condition是数值类型，1就是1，2就是2。#作为输入的condition并不是很合适，一种处理方法就是作为一种向量输入，就是独热编码化。#比如说现在有10个类别，10个类别将被独热编码为长度为10的tensor，使用这个tensor作为我们的condition是比较合适的dataset = torchvision.datasets.MNIST('data',train=True,transform=transform,target_transform=one_hot,download=False)dataloader = data.DataLoader(dataset, batch_size=64, shuffle=True)
这里有个小技巧是作者用到独热编码化
One-Hot编码，又称为一位有效编码，主要是采用位状态寄存器来对个状态进行编码，每个状态都由他独立的寄存器位，并且在任意时候只有一位有效。独热编码 是利用0和1表示一些参数，使用N位状态寄存器来对N个状态进行编码。
例如：参考数字手写体识别中：如数字字体识别0~9中，6的独热编码为
0000001000
二、定义生成器
# 定义生成器class Generator(nn.Module):def __init__(self):super(Generator, self).__init__()self.linear1 = nn.Linear(10, 128 * 7 * 7)self.bn1 = nn.BatchNorm1d(128 * 7 * 7)self.linear2 = nn.Linear(100, 128 * 7 * 7)self.bn2 = nn.BatchNorm1d(128 * 7 * 7)self.deconv1 = nn.ConvTranspose2d(256, 128,kernel_size=(3, 3),padding=1)self.bn3 = nn.BatchNorm2d(128)self.deconv2 = nn.ConvTranspose2d(128, 64,kernel_size=(4, 4),stride=2,padding=1)self.bn4 = nn.BatchNorm2d(64)self.deconv3 = nn.ConvTranspose2d(64, 1,kernel_size=(4, 4),stride=2,padding=1)def forward(self, x1, x2): # X1代表label X2代表imagex1 = F.relu(self.linear1(x1))x1 = self.bn1(x1)x1 = x1.view(-1, 128, 7, 7)x2 = F.relu(self.linear2(x2))x2 = self.bn2(x2)x2 = x2.view(-1, 128, 7, 7)x = torch.cat([x1, x2], axis=1)x = F.relu(self.deconv1(x))x = self.bn3(x)x = F.relu(self.deconv2(x))x = self.bn4(x)x = torch.tanh(self.deconv3(x))return x
三、定义判别器
# 定义判别器# input:1，28，28的图片以及长度为10的conditionclass Discriminator(nn.Module):def __init__(self):super(Discriminator, self).__init__()self.linear = nn.Linear(10, 1*28*28)self.conv1 = nn.Conv2d(2, 64, kernel_size=3, stride=2)self.conv2 = nn.Conv2d(64, 128, kernel_size=3, stride=2)self.bn = nn.BatchNorm2d(128)self.fc = nn.Linear(128*6*6, 1) # 输出一个概率值def forward(self, x1, x2):x1 =F.leaky_relu(self.linear(x1))x1 = x1.view(-1, 1, 28, 28)x = torch.cat([x1, x2], axis=1)x = F.dropout2d(F.leaky_relu(self.conv1(x)))x = F.dropout2d(F.leaky_relu(self.conv2(x)))x = self.bn(x)x = x.view(-1, 128*6*6)x = torch.sigmoid(self.fc(x))return x
四、完整代码展示
import torchimport torch.nn as nnimport torch.nn.functional as Fimport torch.optim as optimimport numpy as npimport matplotlib.pyplot as pltimport torchvisionfrom torchvision import transformsfrom torch.utils import dataimport osimport globfrom PIL import Image# 独热编码# 输入x代表默认的torchvision返回的类比值，class_count类别值为10def one_hot(x, class_count=10):return torch.eye(class_count)[x, :] # 切片选取，第一维选取第x个，第二维全要transform =pose([transforms.ToTensor(),transforms.Normalize(0.5, 0.5)])dataset = torchvision.datasets.MNIST('data',train=True,transform=transform,target_transform=one_hot,download=False)dataloader = data.DataLoader(dataset, batch_size=64, shuffle=True)# 定义生成器class Generator(nn.Module):def __init__(self):super(Generator, self).__init__()self.linear1 = nn.Linear(10, 128 * 7 * 7)self.bn1 = nn.BatchNorm1d(128 * 7 * 7)self.linear2 = nn.Linear(100, 128 * 7 * 7)self.bn2 = nn.BatchNorm1d(128 * 7 * 7)self.deconv1 = nn.ConvTranspose2d(256, 128,kernel_size=(3, 3),padding=1)self.bn3 = nn.BatchNorm2d(128)self.deconv2 = nn.ConvTranspose2d(128, 64,kernel_size=(4, 4),stride=2,padding=1)self.bn4 = nn.BatchNorm2d(64)self.deconv3 = nn.ConvTranspose2d(64, 1,kernel_size=(4, 4),stride=2,padding=1)def forward(self, x1, x2):x1 = F.relu(self.linear1(x1))x1 = self.bn1(x1)x1 = x1.view(-1, 128, 7, 7)x2 = F.relu(self.linear2(x2))x2 = self.bn2(x2)x2 = x2.view(-1, 128, 7, 7)x = torch.cat([x1, x2], axis=1)x = F.relu(self.deconv1(x))x = self.bn3(x)x = F.relu(self.deconv2(x))x = self.bn4(x)x = torch.tanh(self.deconv3(x))return x# 定义判别器# input:1，28，28的图片以及长度为10的conditionclass Discriminator(nn.Module):def __init__(self):super(Discriminator, self).__init__()self.linear = nn.Linear(10, 1*28*28)self.conv1 = nn.Conv2d(2, 64, kernel_size=3, stride=2)self.conv2 = nn.Conv2d(64, 128, kernel_size=3, stride=2)self.bn = nn.BatchNorm2d(128)self.fc = nn.Linear(128*6*6, 1) # 输出一个概率值def forward(self, x1, x2):x1 =F.leaky_relu(self.linear(x1))x1 = x1.view(-1, 1, 28, 28)x = torch.cat([x1, x2], axis=1)x = F.dropout2d(F.leaky_relu(self.conv1(x)))x = F.dropout2d(F.leaky_relu(self.conv2(x)))x = self.bn(x)x = x.view(-1, 128*6*6)x = torch.sigmoid(self.fc(x))return x# 初始化模型device = 'cuda' if torch.cuda.is_available() else 'cpu'gen = Generator().to(device)dis = Discriminator().to(device)# 损失计算函数loss_function = torch.nn.BCELoss()# 定义优化器d_optim = torch.optim.Adam(dis.parameters(), lr=1e-5)g_optim = torch.optim.Adam(gen.parameters(), lr=1e-4)# 定义可视化函数def generate_and_save_images(model, epoch, label_input, noise_input):predictions = np.squeeze(model(label_input, noise_input).cpu().numpy())fig = plt.figure(figsize=(4, 4))for i in range(predictions.shape[0]):plt.subplot(4, 4, i + 1)plt.imshow((predictions[i] + 1) / 2, cmap='gray')plt.axis("off")plt.show()noise_seed = torch.randn(16, 100, device=device)label_seed = torch.randint(0, 10, size=(16,))label_seed_onehot = one_hot(label_seed).to(device)print(label_seed)# print(label_seed_onehot)# 开始训练D_loss = []G_loss = []# 训练循环for epoch in range(150):d_epoch_loss = 0g_epoch_loss = 0count = len(dataloader.dataset)# 对全部的数据集做一次迭代for step, (img, label) in enumerate(dataloader):img = img.to(device)label = label.to(device)size = img.shape[0]random_noise = torch.randn(size, 100, device=device)d_optim.zero_grad()real_output = dis(label, img)d_real_loss = loss_function(real_output,torch.ones_like(real_output, device=device))d_real_loss.backward() #求解梯度# 得到判别器在生成图像上的损失gen_img = gen(label,random_noise)fake_output = dis(label, gen_img.detach()) # 判别器输入生成的图片，f_o是对生成图片的预测结果d_fake_loss = loss_function(fake_output,torch.zeros_like(fake_output, device=device))d_fake_loss.backward()d_loss = d_real_loss + d_fake_lossd_optim.step() # 优化# 得到生成器的损失g_optim.zero_grad()fake_output = dis(label, gen_img)g_loss = loss_function(fake_output,torch.ones_like(fake_output, device=device))g_loss.backward()g_optim.step()with torch.no_grad():d_epoch_loss += d_loss.item()g_epoch_loss += g_loss.item()with torch.no_grad():d_epoch_loss /= countg_epoch_loss /= countD_loss.append(d_epoch_loss)G_loss.append(g_epoch_loss)if epoch % 10 == 0:print('Epoch:', epoch)generate_and_save_images(gen, epoch, label_seed_onehot, noise_seed)
五、结果展示
随机生成的条件
根据这个条件生成的手写数字
