GAN实战之Pytorch 使用CGAN生成指定MNIST手写数字
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 =transforms.Compose([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image
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condition
class 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 torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
import numpy as np
import matplotlib.pyplot as plt
import torchvision
from torchvision import transforms
from torch.utils import data
import os
import glob
from 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10
def 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 =transforms.Compose([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condition
class 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 = 0
g_epoch_loss = 0
count = 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_loss
d_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 /= count
g_epoch_loss /= count
D_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
