ACGAN代码——阅读记录

步骤

定义初始化参数
定义生成器网络（标签 尺寸）-将标签和噪声作为输入-全连接层-重排数据-卷积块-生成图片
定义判别器网络 -将生成图片输入卷积块-重排数据—1.输出真伪2.输出分类
初始化参数
数据加载
定义优化器
定义采样
训练: G用随机噪声和标签生成图片——用D打分——定义g_loss 优化G

import argparse
import os
import numpy as np
import math

import torchvision.transforms as transforms
from torchvision.utils import save_image

from torch.utils.data import DataLoader
from torchvision import datasets
from torch.autograd import Variable

import torch.nn as nn
import torch.nn.functional as F
import torch

os.makedirs("images", exist_ok=True)

parser = argparse.ArgumentParser()
parser.add_argument("--n_epochs", type=int, default=200, help="number of epochs of training")
parser.add_argument("--batch_size", type=int, default=64, help="size of the batches")
parser.add_argument("--lr", type=float, default=0.0002, help="adam: learning rate")
parser.add_argument("--b1", type=float, default=0.5, help="adam: decay of first order momentum of gradient")
parser.add_argument("--b2", type=float, default=0.999, help="adam: decay of first order momentum of gradient")
parser.add_argument("--n_cpu", type=int, default=8, help="number of cpu threads to use during batch generation")
parser.add_argument("--latent_dim", type=int, default=100, help="dimensionality of the latent space")
parser.add_argument("--n_classes", type=int, default=10, help="number of classes for dataset")  #类别数量
parser.add_argument("--img_size", type=int, default=32, help="size of each image dimension")        #图片尺寸
parser.add_argument("--channels", type=int, default=1, help="number of image channels")             #彩色/黑白
parser.add_argument("--sample_interval", type=int, default=400, help="interval between image sampling") #采样频率
opt = parser.parse_args()
print(opt)

cuda = True if torch.cuda.is_available() else False


def weights_init_normal(m):     #定义参数初始化函数，解析：https://www.lizenghai.com/archives/28867.html
    classname = m.__class__.__name__
    if classname.find("Conv") != -1:
        torch.nn.init.normal_(m.weight.data, 0.0, 0.02)
    elif classname.find("BatchNorm2d") != -1:
        torch.nn.init.normal_(m.weight.data, 1.0, 0.02)
        torch.nn.init.constant_(m.bias.data, 0.0)


class Generator(nn.Module):
    def __init__(self):
        super(Generator, self).__init__()

        self.label_emb = nn.Embedding(opt.n_classes, opt.latent_dim)    #词嵌入(类别数，维度)  ACGAN的条件信息

        self.init_size = opt.img_size // 4  # Initial size before upsampling 整除
        self.l1 = nn.Sequential(nn.Linear(opt.latent_dim, 128 * self.init_size ** 2))   #输入100维噪声  全连接层

        self.conv_blocks = nn.Sequential(       #CNN用于GAN中要变动
            nn.BatchNorm2d(128),    #卷积的输出通道数 128  （理解为128个卷积核 128片）
            nn.Upsample(scale_factor=2),    #上采样  指定输出为输入的多少倍数
            nn.Conv2d(128, 128, 3, stride=1, padding=1),    #nn.Conv2d(in_channel, out_channel, 3, stride, 1, bias=False)
            nn.BatchNorm2d(128, 0.8),  #0.8为使数值稳定而加到分母上的值
            nn.LeakyReLU(0.2, inplace=True),    #0.2=控制负斜率的角度，inplace-选择是否进行覆盖运算
            nn.Upsample(scale_factor=2),
            nn.Conv2d(128, 64, 3, stride=1, padding=1),
            nn.BatchNorm2d(64, 0.8),
            nn.LeakyReLU(0.2, inplace=True),
            nn.Conv2d(64, opt.channels, 3, stride=1, padding=1),
            nn.Tanh(),
        )

    def forward(self, noise, labels):
        gen_input = torch.mul(self.label_emb(labels), noise)    #torch.mul(input, value, out=None)
        out = self.l1(gen_input)        #经过l1网络全连接层
        out = out.view(out.shape[0], 128, self.init_size, self.init_size)   #通道数
        img = self.conv_blocks(out)     #卷积块
        return img


class Discriminator(nn.Module):
    def __init__(self):
        super(Discriminator, self).__init__()

        def discriminator_block(in_filters, out_filters, bn=True):
            """Returns layers of each discriminator block"""
            block = [nn.Conv2d(in_filters, out_filters, 3, 2, 1), nn.LeakyReLU(0.2, inplace=True), nn.Dropout2d(0.25)]
            if bn:
                block.append(nn.BatchNorm2d(out_filters, 0.8))
            return block

        self.conv_blocks = nn.Sequential(
            *discriminator_block(opt.channels, 16, bn=False), #discriminator_block在81行
            *discriminator_block(16, 32),
            *discriminator_block(32, 64),
            *discriminator_block(64, 128),
        )

        # The height and width of downsampled image
        ds_size = opt.img_size // 2 ** 4

        # Output layers
        self.adv_layer = nn.Sequential(nn.Linear(128 * ds_size ** 2, 1), nn.Sigmoid())
        self.aux_layer = nn.Sequential(nn.Linear(128 * ds_size ** 2, opt.n_classes), nn.Softmax())

    def forward(self, img):
        out = self.conv_blocks(img) #卷积块
        out = out.view(out.shape[0], -1)    #平铺
        validity = self.adv_layer(out)      #对抗的结果
        label = self.aux_layer(out)         #分类的结果

        return validity, label


# Loss functions
adversarial_loss = torch.nn.BCELoss()
auxiliary_loss = torch.nn.CrossEntropyLoss()

# Initialize generator and discriminator
generator = Generator()
discriminator = Discriminator()

if cuda:
    generator.cuda()
    discriminator.cuda()
    adversarial_loss.cuda()
    auxiliary_loss.cuda()

# Initialize weights
generator.apply(weights_init_normal)
discriminator.apply(weights_init_normal)

# Configure data loader
os.makedirs("../../data/mnist", exist_ok=True)
dataloader = torch.utils.data.DataLoader(
    datasets.MNIST(
        "../../data/mnist",
        train=True,
        download=True,
        transform=transforms.Compose(
            [transforms.Resize(opt.img_size), transforms.ToTensor(), transforms.Normalize([0.5], [0.5])]
        ),
    ),
    batch_size=opt.batch_size,
    shuffle=True,
)

# Optimizers
optimizer_G = torch.optim.Adam(generator.parameters(), lr=opt.lr, betas=(opt.b1, opt.b2))
optimizer_D = torch.optim.Adam(discriminator.parameters(), lr=opt.lr, betas=(opt.b1, opt.b2))

FloatTensor = torch.cuda.FloatTensor if cuda else torch.FloatTensor
LongTensor = torch.cuda.LongTensor if cuda else torch.LongTensor


def sample_image(n_row, batches_done):
    """Saves a grid of generated digits ranging from 0 to n_classes"""
    # Sample noise
    z = FloatTensor(np.random.normal(0, 1, (n_row ** 2, opt.latent_dim)))
    # Get labels ranging from 0 to n_classes for n rows
    labels = np.array([num for _ in range(n_row) for num in range(n_row)])
    labels = LongTensor(labels)#放入GPU
    gen_imgs = generator(z, labels) #生成图像有标签
    save_image(gen_imgs.data, "images/%d.png" % batches_done, nrow=n_row, normalize=True)   #保存到文件夹


# ----------
#  Training
# ----------

for epoch in range(opt.n_epochs):
    for i, (imgs, labels) in enumerate(dataloader):

        batch_size = imgs.shape[0]

        # Adversarial ground truths   放入cuda
        valid = Variable(FloatTensor(batch_size, 1).fill_(1.0), requires_grad=False)
        fake = Variable(FloatTensor(batch_size, 1).fill_(0.0), requires_grad=False)
        #valid和fake代表不同的值 训练D时让真实数据值和valid接近  伪造数据值和fake接近
        # Configure input
        real_imgs = imgs.type(FloatTensor)
        labels = labels.type(LongTensor)

        # -----------------
        #  Train Generator 先训练生成器 得到标签
        # -----------------

        optimizer_G.zero_grad()

        # Sample noise and labels as generator input
        z = Variable(FloatTensor(np.random.normal(0, 1, (batch_size, opt.latent_dim)))) #均值、标准差、shape（一批多少个z 一个z多少维）
        gen_labels = Variable(LongTensor(np.random.randint(0, opt.n_classes, batch_size)))# 返回随机整数[low,high),多少个

        # Generate a batch of images
        gen_imgs = generator(z, gen_labels)

        # Loss measures generator's ability to fool the discriminator
        validity, pred_label = discriminator(gen_imgs)
        g_loss = 0.5 * (adversarial_loss(validity, valid) + auxiliary_loss(pred_label, gen_labels))
        # 判别器给gen_img的得分和预标签为validity, pred_label    生成器的输入为z,gen_labels    希望真实数据的值为valid
        g_loss.backward()
        optimizer_G.step()

        # ---------------------
        #  Train Discriminator
        # ---------------------

        optimizer_D.zero_grad()
        #鉴别器loss得定义
        # Loss for real images
        real_pred, real_aux = discriminator(real_imgs)  #真实数据评分和标签
        d_real_loss = (adversarial_loss(real_pred, valid) + auxiliary_loss(real_aux, labels)) / 2
                            #要让valid和real_pred接近    并且让label（直接取得）和判别的aux一致
        # Loss for fake images
        fake_pred, fake_aux = discriminator(gen_imgs.detach())
        d_fake_loss = (adversarial_loss(fake_pred, fake) + auxiliary_loss(fake_aux, gen_labels)) / 2
                            # 要让fake和fake_pred接近    并且让gen_label（直接取得）和判别的aux一致
        # Total discriminator loss
        d_loss = (d_real_loss + d_fake_loss) / 2

        # Calculate discriminator accuracy   concatenate数组拼接
        pred = np.concatenate([real_aux.data.cpu().numpy(), fake_aux.data.cpu().numpy()], axis=0)#鉴别器判别的
        gt = np.concatenate([labels.data.cpu().numpy(), gen_labels.data.cpu().numpy()], axis=0)     #实际取得
        d_acc = np.mean(np.argmax(pred, axis=1) == gt)      #判别和取得 相同程度

        d_loss.backward()
        optimizer_D.step()

        print(
            "[Epoch %d/%d] [Batch %d/%d] [D loss: %f, acc: %d%%] [G loss: %f]"
            % (epoch, opt.n_epochs, i, len(dataloader), d_loss.item(), 100 * d_acc, g_loss.item())
        )
        batches_done = epoch * len(dataloader) + i
        if batches_done % opt.sample_interval == 0:
            sample_image(n_row=10, batches_done=batches_done)