1. DCGAN简单总结
   DCGAN的全称是Deep Convolutional Generative Adversarial Networks ,
   意即深度卷积对抗生成网络，它是由Alec Radford在论文Unsupervised
   Representation Learning with Deep Convolutional Generative Adversarial
   Networks中提出的。从名字上来看，它是在GAN的基础上增加深度卷积网
   络结构，专门生成图像样本。
   在GAN中 做了D 、G 的输入输出租损失的走义，但事实上， GAN 并没再对D 、G 的具体结构做出任何限制。DCGAN中的D 、G 的含义以及损失都和原始GAN中完全一致，但是它在D和G中采用了较为特殊的结构，以便对图片进行有效建模。对于判别器D，它的输入是一张图像，输出是这张图像为真实图像的概率。在DCGAN中，判别器D的结构是一个卷积神经网络，输入的图像经过若干层卷积后得到一个卷积特征，将得到的特征送入Logistic函数，输出可以看作是概率。
   稳定的深度卷积GAN 架构指南:

• 所有的pooling层使用步幅卷积(判别网络)和微步幅度卷积(生成网络)进行替换。
• 在生成网络和判别网络上使用批处理规范化。
• 在生成网络和判别网络上使用批处理规范化。
• 对于更深的架构移除全连接隐藏层。
• 在生成网络的所有层上使用RelU激活函数，除了输出层使用Tanh激活函数。
• 在判别网络的所有层上使用LeakyReLU激活函数。

2. 搜集原始数据集
   首先我们需要用爬虫爬取大量的动漫图片，从著名的动漫图库网站爬取：http://konachan.net/

源代码如下：

import requests  # http lib
from bs4 import BeautifulSoup  # climb lib
import os # operation system
import traceback # trace deviance

def download(url,filename):
    if os.path.exists(filename):
        print('file exists!')
        return
    try:
        r = requests.get(url,stream=True,timeout=60)
        r.raise_for_status()
        with open(filename,'wb') as f:
            for chunk in r.iter_content(chunk_size=1024):
                if chunk: # filter out keep-alove new chunks
                    f.write(chunk)
                    f.flush()
        return filename
    except KeyboardInterrupt:
        if os.path.exists(filename):
            os.remove(filename)
        return KeyboardInterrupt
    except Exception:
        traceback.print_exc()
        if os.path.exists(filename):
            os.remove(filename)

if os.path.exists('imgs') is False:
    os.makedirs('imgs')

start = 1
end = 8000
for i in range(start, end+1):
    url = 'http://konachan.net/post?page=%d&tags=' % i
    html = requests.get(url).text # gain the web's information
    soup =  BeautifulSoup(html,'html.parser') # doc's string and jie xi qi
    for img in soup.find_all('img',class_="preview"):# 遍历所有preview类，找到img标签
        target_url = img['src']
        filename = os.path.join('imgs',target_url.split('/')[-1])
        download(target_url,filename)
    print('%d / %d' % (i,end))    

下载的部分图像如下：

现在已经有了基本的图像了，但我们的目标是生成动漫头像，不需要整张图像，而且其他的信息会干扰到训练，所以需要进行人脸检测截取人脸图像。

3. 人脸检测截取人脸

通过基于opencv的人脸检测分类器，参考于lbpcascade_animeface。

首先，要使用这个分类器要先进行下载：

wget https://raw.githubusercontent.com/nagadomi/lbpcascade_animeface/master/lbpcascade_animeface.xml

下载完成后，运行以下代码对图像进行人脸截取。

import cv2
import sys
import os.path
from glob import glob

def detect(filename,cascade_file="lbpcascade_animeface.xml"):
    if not os.path.isfile(cascade_file):
        raise RuntimeError("%s: not found" % cascade_file)

    cascade = cv2.CascadeClassifier(cascade_file)
    image = cv2.imread(filename)
    gray = cv2.cvtColor(image,cv2.COLOR_BGR2GRAY)
    gray = cv2.equalizeHist(gray)

    faces = cascade.detectMultiScale(
        gray,
        # detector options
        scaleFactor = 1.1,
        minNeighbors = 5,
        minSize = (48,48)
    )

    for i,(x,y,w,h) in enumerate(faces):
        face = image[y: y+h, x:x+w, :]
        face = cv2.resize(face,(96,96))
        save_filename = '%s.jpg' % (os.path.basename(filename).split('.')[0])
        cv2.imwrite("faces/"+save_filename,face)

if __name__ == '__main__':
    if os.path.exists('faces') is False:
        os.makedirs('faces')
    file_list = glob('imgs/*.jpg')
    for filename in file_list:
        detect(filename)

处理后的图像如下所示：

4. 源代码分析
   参照于DCGAN-tensorflow

总共获取18,466张图像，人脸检测后得到5,238张。

需要安装tensorflow，共4个文件，分别是main.py、model.py、ops.py、utils.py。

• model.py

from __future__ import division
import os
import time
import math
from glob import glob
import tensorflow as tf
import numpy as np
from six.moves import xrange

from ops import *
from utils import *

def conv_out_size_same(size, stride):
  return int(math.ceil(float(size) / float(stride)))

class DCGAN(object):
  def __init__(self, sess, input_height=108, input_width=108, crop=True,
         batch_size=64, sample_num = 64, output_height=64, output_width=64,
         y_dim=None, z_dim=100, gf_dim=64, df_dim=64,
         gfc_dim=1024, dfc_dim=1024, c_dim=3, dataset_name='default',
         input_fname_pattern='*.jpg', checkpoint_dir=None, sample_dir=None):
    """

    Args:
      sess: TensorFlow session
      batch_size: The size of batch. Should be specified before training.
      y_dim: (optional) Dimension of dim for y. [None]
      z_dim: (optional) Dimension of dim for Z. [100]
      gf_dim: (optional) Dimension of gen filters in first conv layer. [64]
      df_dim: (optional) Dimension of discrim filters in first conv layer. [64]
      gfc_dim: (optional) Dimension of gen units for for fully connected layer. [1024]
      dfc_dim: (optional) Dimension of discrim units for fully connected layer. [1024]
      c_dim: (optional) Dimension of image color. For grayscale input, set to 1. [3]
    """
    self.sess = sess
    self.crop = crop

    self.batch_size = batch_size
    self.sample_num = sample_num

    self.input_height = input_height
    self.input_width = input_width
    self.output_height = output_height
    self.output_width = output_width

    self.y_dim = y_dim
    self.z_dim = z_dim

    self.gf_dim = gf_dim
    self.df_dim = df_dim

    self.gfc_dim = gfc_dim
    self.dfc_dim = dfc_dim

    # batch normalization : deals with poor initialization helps gradient flow
    self.d_bn1 = batch_norm(name='d_bn1')
    self.d_bn2 = batch_norm(name='d_bn2')

    if not self.y_dim:
      self.d_bn3 = batch_norm(name='d_bn3')

    self.g_bn0 = batch_norm(name='g_bn0')
    self.g_bn1 = batch_norm(name='g_bn1')
    self.g_bn2 = batch_norm(name='g_bn2')

    if not self.y_dim:
      self.g_bn3 = batch_norm(name='g_bn3')

    self.dataset_name = dataset_name
    self.input_fname_pattern = input_fname_pattern
    self.checkpoint_dir = checkpoint_dir

    if self.dataset_name == 'mnist':
      self.data_X, self.data_y = self.load_mnist()
      self.c_dim = self.data_X[0].shape[-1]
    else:
      self.data = glob(os.path.join("./data", self.dataset_name, self.input_fname_pattern))
      imreadImg = imread(self.data[0]);
      if len(imreadImg.shape) >= 3: #check if image is a non-grayscale image by checking channel number
        self.c_dim = imread(self.data[0]).shape[-1]
      else:
        self.c_dim = 1

    self.grayscale = (self.c_dim == 1)

    self.build_model()

  def build_model(self):
    if self.y_dim:
      self.y= tf.placeholder(tf.float32, [self.batch_size, self.y_dim], name='y')

    if self.crop:
      image_dims = [self.output_height, self.output_width, self.c_dim]
    else:
      image_dims = [self.input_height, self.input_width, self.c_dim]

    self.inputs = tf.placeholder(
      tf.float32, [self.batch_size] + image_dims, name='real_images')

    inputs = self.inputs

    self.z = tf.placeholder(
      tf.float32, [None, self.z_dim], name='z')
    self.z_sum = histogram_summary("z", self.z)

    if self.y_dim:
      self.G = self.generator(self.z, self.y)
      self.D, self.D_logits = \
          self.discriminator(inputs, self.y, reuse=False)

      self.sampler = self.sampler(self.z, self.y)
      self.D_, self.D_logits_ = \
          self.discriminator(self.G, self.y, reuse=True)
    else:
      self.G = self.generator(self.z)
      self.D, self.D_logits = self.discriminator(inputs)

      self.sampler = self.sampler(self.z)
      self.D_, self.D_logits_ = self.discriminator(self.G, reuse=True)

    self.d_sum = histogram_summary("d", self.D)
    self.d__sum = histogram_summary("d_", self.D_)
    self.G_sum = image_summary("G", self.G)

    def sigmoid_cross_entropy_with_logits(x, y):
      try:
        return tf.nn.sigmoid_cross_entropy_with_logits(logits=x, labels=y)
      except:
        return tf.nn.sigmoid_cross_entropy_with_logits(logits=x, targets=y)

    self.d_loss_real = tf.reduce_mean(
      sigmoid_cross_entropy_with_logits(self.D_logits, tf.ones_like(self.D)))
    self.d_loss_fake = tf.reduce_mean(
      sigmoid_cross_entropy_with_logits(self.D_logits_, tf.zeros_like(self.D_)))
    self.g_loss = tf.reduce_mean(
      sigmoid_cross_entropy_with_logits(self.D_logits_, tf.ones_like(self.D_)))

    self.d_loss_real_sum = scalar_summary("d_loss_real", self.d_loss_real)
    self.d_loss_fake_sum = scalar_summary("d_loss_fake", self.d_loss_fake)
                          
    self.d_loss = self.d_loss_real + self.d_loss_fake

    self.g_loss_sum = scalar_summary("g_loss", self.g_loss)
    self.d_loss_sum = scalar_summary("d_loss", self.d_loss)

    t_vars = tf.trainable_variables()

    self.d_vars = [var for var in t_vars if 'd_' in var.name]
    self.g_vars = [var for var in t_vars if 'g_' in var.name]

    self.saver = tf.train.Saver()

  def train(self, config):
    d_optim = tf.train.AdamOptimizer(config.learning_rate, beta1=config.beta1) \
              .minimize(self.d_loss, var_list=self.d_vars)
    g_optim = tf.train.AdamOptimizer(config.learning_rate, beta1=config.beta1) \
              .minimize(self.g_loss, var_list=self.g_vars)
    try:
      tf.global_variables_initializer().run()
    except:
      tf.initialize_all_variables().run()

    self.g_sum = merge_summary([self.z_sum, self.d__sum,
      self.G_sum, self.d_loss_fake_sum, self.g_loss_sum])
    self.d_sum = merge_summary(
        [self.z_sum, self.d_sum, self.d_loss_real_sum, self.d_loss_sum])
    self.writer = SummaryWriter("./logs", self.sess.graph)

    sample_z = np.random.uniform(-1, 1, size=(self.sample_num , self.z_dim))
    
    if config.dataset == 'mnist':
      sample_inputs = self.data_X[0:self.sample_num]
      sample_labels = self.data_y[0:self.sample_num]
    else:
      sample_files = self.data[0:self.sample_num]
      sample = [
          get_image(sample_file,
                    input_height=self.input_height,
                    input_width=self.input_width,
                    resize_height=self.output_height,
                    resize_width=self.output_width,
                    crop=self.crop,
                    grayscale=self.grayscale) for sample_file in sample_files]
      if (self.grayscale):
        sample_inputs = np.array(sample).astype(np.float32)[:, :, :, None]
      else:
        sample_inputs = np.array(sample).astype(np.float32)
  
    counter = 1
    start_time = time.time()
    could_load, checkpoint_counter = self.load(self.checkpoint_dir)
    if could_load:
      counter = checkpoint_counter
      print(" [*] Load SUCCESS")
    else:
      print(" [!] Load failed...")

    for epoch in xrange(config.epoch):
      if config.dataset == 'mnist':
        batch_idxs = min(len(self.data_X), config.train_size) // config.batch_size
      else:      
        self.data = glob(os.path.join(
          "./data", config.dataset, self.input_fname_pattern))
        batch_idxs = min(len(self.data), config.train_size) // config.batch_size

      for idx in xrange(0, batch_idxs):
        if config.dataset == 'mnist':
          batch_images = self.data_X[idx*config.batch_size:(idx+1)*config.batch_size]
          batch_labels = self.data_y[idx*config.batch_size:(idx+1)*config.batch_size]
        else:
          batch_files = self.data[idx*config.batch_size:(idx+1)*config.batch_size]
          batch = [
              get_image(batch_file,
                        input_height=self.input_height,
                        input_width=self.input_width,
                        resize_height=self.output_height,
                        resize_width=self.output_width,
                        crop=self.crop,
                        grayscale=self.grayscale) for batch_file in batch_files]
          if self.grayscale:
            batch_images = np.array(batch).astype(np.float32)[:, :, :, None]
          else:
            batch_images = np.array(batch).astype(np.float32)

        batch_z = np.random.uniform(-1, 1, [config.batch_size, self.z_dim]) \
              .astype(np.float32)

        if config.dataset == 'mnist':
          # Update D network
          _, summary_str = self.sess.run([d_optim, self.d_sum],
            feed_dict={ 
              self.inputs: batch_images,
              self.z: batch_z,
              self.y:batch_labels,
            })
          self.writer.add_summary(summary_str, counter)

          # Update G network
          _, summary_str = self.sess.run([g_optim, self.g_sum],
            feed_dict={
              self.z: batch_z, 
              self.y:batch_labels,
            })
          self.writer.add_summary(summary_str, counter)

          # Run g_optim twice to make sure that d_loss does not go to zero (different from paper)
          _, summary_str = self.sess.run([g_optim, self.g_sum],
            feed_dict={ self.z: batch_z, self.y:batch_labels })
          self.writer.add_summary(summary_str, counter)
          
          errD_fake = self.d_loss_fake.eval({
              self.z: batch_z, 
              self.y:batch_labels
          })
          errD_real = self.d_loss_real.eval({
              self.inputs: batch_images,
              self.y:batch_labels
          })
          errG = self.g_loss.eval({
              self.z: batch_z,
              self.y: batch_labels
          })
        else:
          # Update D network
          _, summary_str = self.sess.run([d_optim, self.d_sum],
            feed_dict={ self.inputs: batch_images, self.z: batch_z })
          self.writer.add_summary(summary_str, counter)

          # Update G network
          _, summary_str = self.sess.run([g_optim, self.g_sum],
            feed_dict={ self.z: batch_z })
          self.writer.add_summary(summary_str, counter)

          # Run g_optim twice to make sure that d_loss does not go to zero (different from paper)
          _, summary_str = self.sess.run([g_optim, self.g_sum],
            feed_dict={ self.z: batch_z })
          self.writer.add_summary(summary_str, counter)
          
          errD_fake = self.d_loss_fake.eval({ self.z: batch_z })
          errD_real = self.d_loss_real.eval({ self.inputs: batch_images })
          errG = self.g_loss.eval({self.z: batch_z})

        counter += 1
        print("Epoch: [%2d] [%4d/%4d] time: %4.4f, d_loss: %.8f, g_loss: %.8f" \
          % (epoch, idx, batch_idxs,
            time.time() - start_time, errD_fake+errD_real, errG))

        if np.mod(counter, 100) == 1:
          if config.dataset == 'mnist':
            samples, d_loss, g_loss = self.sess.run(
              [self.sampler, self.d_loss, self.g_loss],
              feed_dict={
                  self.z: sample_z,
                  self.inputs: sample_inputs,
                  self.y:sample_labels,
              }
            )
            save_images(samples, image_manifold_size(samples.shape[0]),
                  './{}/train_{:02d}_{:04d}.png'.format(config.sample_dir, epoch, idx))
            print("[Sample] d_loss: %.8f, g_loss: %.8f" % (d_loss, g_loss)) 
          else:
            try:
              samples, d_loss, g_loss = self.sess.run(
                [self.sampler, self.d_loss, self.g_loss],
                feed_dict={
                    self.z: sample_z,
                    self.inputs: sample_inputs,
                },
              )
              save_images(samples, image_manifold_size(samples.shape[0]),
                    './{}/train_{:02d}_{:04d}.png'.format(config.sample_dir, epoch, idx))
              print("[Sample] d_loss: %.8f, g_loss: %.8f" % (d_loss, g_loss)) 
            except:
              print("one pic error!...")

        if np.mod(counter, 500) == 2:
          self.save(config.checkpoint_dir, counter)

  def discriminator(self, image, y=None, reuse=False):
    with tf.variable_scope("discriminator") as scope:
      if reuse:
        scope.reuse_variables()

      if not self.y_dim:
        h0 = lrelu(conv2d(image, self.df_dim, name='d_h0_conv'))
        h1 = lrelu(self.d_bn1(conv2d(h0, self.df_dim*2, name='d_h1_conv')))
        h2 = lrelu(self.d_bn2(conv2d(h1, self.df_dim*4, name='d_h2_conv')))
        h3 = lrelu(self.d_bn3(conv2d(h2, self.df_dim*8, name='d_h3_conv')))
        h4 = linear(tf.reshape(h3, [self.batch_size, -1]), 1, 'd_h4_lin')

        return tf.nn.sigmoid(h4), h4
      else:
        yb = tf.reshape(y, [self.batch_size, 1, 1, self.y_dim])
        x = conv_cond_concat(image, yb)

        h0 = lrelu(conv2d(x, self.c_dim + self.y_dim, name='d_h0_conv'))
        h0 = conv_cond_concat(h0, yb)

        h1 = lrelu(self.d_bn1(conv2d(h0, self.df_dim + self.y_dim, name='d_h1_conv')))
        h1 = tf.reshape(h1, [self.batch_size, -1])      
        h1 = concat([h1, y], 1)
        
        h2 = lrelu(self.d_bn2(linear(h1, self.dfc_dim, 'd_h2_lin')))
        h2 = concat([h2, y], 1)

        h3 = linear(h2, 1, 'd_h3_lin')
        
        return tf.nn.sigmoid(h3), h3

  def generator(self, z, y=None):
    with tf.variable_scope("generator") as scope:
      if not self.y_dim:
        s_h, s_w = self.output_height, self.output_width
        s_h2, s_w2 = conv_out_size_same(s_h, 2), conv_out_size_same(s_w, 2)
        s_h4, s_w4 = conv_out_size_same(s_h2, 2), conv_out_size_same(s_w2, 2)
        s_h8, s_w8 = conv_out_size_same(s_h4, 2), conv_out_size_same(s_w4, 2)
        s_h16, s_w16 = conv_out_size_same(s_h8, 2), conv_out_size_same(s_w8, 2)

        # project `z` and reshape
        self.z_, self.h0_w, self.h0_b = linear(
            z, self.gf_dim*8*s_h16*s_w16, 'g_h0_lin', with_w=True)

        self.h0 = tf.reshape(
            self.z_, [-1, s_h16, s_w16, self.gf_dim * 8])
        h0 = tf.nn.relu(self.g_bn0(self.h0))

        self.h1, self.h1_w, self.h1_b = deconv2d(
            h0, [self.batch_size, s_h8, s_w8, self.gf_dim*4], name='g_h1', with_w=True)
        h1 = tf.nn.relu(self.g_bn1(self.h1))

        h2, self.h2_w, self.h2_b = deconv2d(
            h1, [self.batch_size, s_h4, s_w4, self.gf_dim*2], name='g_h2', with_w=True)
        h2 = tf.nn.relu(self.g_bn2(h2))

        h3, self.h3_w, self.h3_b = deconv2d(
            h2, [self.batch_size, s_h2, s_w2, self.gf_dim*1], name='g_h3', with_w=True)
        h3 = tf.nn.relu(self.g_bn3(h3))

        h4, self.h4_w, self.h4_b = deconv2d(
            h3, [self.batch_size, s_h, s_w, self.c_dim], name='g_h4', with_w=True)

        return tf.nn.tanh(h4)
      else:
        s_h, s_w = self.output_height, self.output_width
        s_h2, s_h4 = int(s_h/2), int(s_h/4)
        s_w2, s_w4 = int(s_w/2), int(s_w/4)

        # yb = tf.expand_dims(tf.expand_dims(y, 1),2)
        yb = tf.reshape(y, [self.batch_size, 1, 1, self.y_dim])
        z = concat([z, y], 1)

        h0 = tf.nn.relu(
            self.g_bn0(linear(z, self.gfc_dim, 'g_h0_lin')))
        h0 = concat([h0, y], 1)

        h1 = tf.nn.relu(self.g_bn1(
            linear(h0, self.gf_dim*2*s_h4*s_w4, 'g_h1_lin')))
        h1 = tf.reshape(h1, [self.batch_size, s_h4, s_w4, self.gf_dim * 2])

        h1 = conv_cond_concat(h1, yb)

        h2 = tf.nn.relu(self.g_bn2(deconv2d(h1,
            [self.batch_size, s_h2, s_w2, self.gf_dim * 2], name='g_h2')))
        h2 = conv_cond_concat(h2, yb)

        return tf.nn.sigmoid(
            deconv2d(h2, [self.batch_size, s_h, s_w, self.c_dim], name='g_h3'))

  def sampler(self, z, y=None):
    with tf.variable_scope("generator") as scope:
      scope.reuse_variables()

      if not self.y_dim:
        s_h, s_w = self.output_height, self.output_width
        s_h2, s_w2 = conv_out_size_same(s_h, 2), conv_out_size_same(s_w, 2)
        s_h4, s_w4 = conv_out_size_same(s_h2, 2), conv_out_size_same(s_w2, 2)
        s_h8, s_w8 = conv_out_size_same(s_h4, 2), conv_out_size_same(s_w4, 2)
        s_h16, s_w16 = conv_out_size_same(s_h8, 2), conv_out_size_same(s_w8, 2)

        # project `z` and reshape
        h0 = tf.reshape(
            linear(z, self.gf_dim*8*s_h16*s_w16, 'g_h0_lin'),
            [-1, s_h16, s_w16, self.gf_dim * 8])
        h0 = tf.nn.relu(self.g_bn0(h0, train=False))

        h1 = deconv2d(h0, [self.batch_size, s_h8, s_w8, self.gf_dim*4], name='g_h1')
        h1 = tf.nn.relu(self.g_bn1(h1, train=False))

        h2 = deconv2d(h1, [self.batch_size, s_h4, s_w4, self.gf_dim*2], name='g_h2')
        h2 = tf.nn.relu(self.g_bn2(h2, train=False))

        h3 = deconv2d(h2, [self.batch_size, s_h2, s_w2, self.gf_dim*1], name='g_h3')
        h3 = tf.nn.relu(self.g_bn3(h3, train=False))

        h4 = deconv2d(h3, [self.batch_size, s_h, s_w, self.c_dim], name='g_h4')

        return tf.nn.tanh(h4)
      else:
        s_h, s_w = self.output_height, self.output_width
        s_h2, s_h4 = int(s_h/2), int(s_h/4)
        s_w2, s_w4 = int(s_w/2), int(s_w/4)

        # yb = tf.reshape(y, [-1, 1, 1, self.y_dim])
        yb = tf.reshape(y, [self.batch_size, 1, 1, self.y_dim])
        z = concat([z, y], 1)

        h0 = tf.nn.relu(self.g_bn0(linear(z, self.gfc_dim, 'g_h0_lin'), train=False))
        h0 = concat([h0, y], 1)

        h1 = tf.nn.relu(self.g_bn1(
            linear(h0, self.gf_dim*2*s_h4*s_w4, 'g_h1_lin'), train=False))
        h1 = tf.reshape(h1, [self.batch_size, s_h4, s_w4, self.gf_dim * 2])
        h1 = conv_cond_concat(h1, yb)

        h2 = tf.nn.relu(self.g_bn2(
            deconv2d(h1, [self.batch_size, s_h2, s_w2, self.gf_dim * 2], name='g_h2'), train=False))
        h2 = conv_cond_concat(h2, yb)

        return tf.nn.sigmoid(deconv2d(h2, [self.batch_size, s_h, s_w, self.c_dim], name='g_h3'))

  def load_mnist(self):
    data_dir = os.path.join("./data", self.dataset_name)
    
    fd = open(os.path.join(data_dir,'train-images-idx3-ubyte'))
    loaded = np.fromfile(file=fd,dtype=np.uint8)
    trX = loaded[16:].reshape((60000,28,28,1)).astype(np.float)

    fd = open(os.path.join(data_dir,'train-labels-idx1-ubyte'))
    loaded = np.fromfile(file=fd,dtype=np.uint8)
    trY = loaded[8:].reshape((60000)).astype(np.float)

    fd = open(os.path.join(data_dir,'t10k-images-idx3-ubyte'))
    loaded = np.fromfile(file=fd,dtype=np.uint8)
    teX = loaded[16:].reshape((10000,28,28,1)).astype(np.float)

    fd = open(os.path.join(data_dir,'t10k-labels-idx1-ubyte'))
    loaded = np.fromfile(file=fd,dtype=np.uint8)
    teY = loaded[8:].reshape((10000)).astype(np.float)

    trY = np.asarray(trY)
    teY = np.asarray(teY)
    
    X = np.concatenate((trX, teX), axis=0)
    y = np.concatenate((trY, teY), axis=0).astype(np.int)
    
    seed = 547
    np.random.seed(seed)
    np.random.shuffle(X)
    np.random.seed(seed)
    np.random.shuffle(y)
    
    y_vec = np.zeros((len(y), self.y_dim), dtype=np.float)
    for i, label in enumerate(y):
      y_vec[i,y[i]] = 1.0
    
    return X/255.,y_vec

  @property
  def model_dir(self):
    return "{}_{}_{}_{}".format(
        self.dataset_name, self.batch_size,
        self.output_height, self.output_width)
      
  def save(self, checkpoint_dir, step):
    model_name = "DCGAN.model"
    checkpoint_dir = os.path.join(checkpoint_dir, self.model_dir)

    if not os.path.exists(checkpoint_dir):
      os.makedirs(checkpoint_dir)

    self.saver.save(self.sess,
            os.path.join(checkpoint_dir, model_name),
            global_step=step)

  def load(self, checkpoint_dir):
    import re
    print(" [*] Reading checkpoints...")
    checkpoint_dir = os.path.join(checkpoint_dir, self.model_dir)

    ckpt = tf.train.get_checkpoint_state(checkpoint_dir)
    if ckpt and ckpt.model_checkpoint_path:
      ckpt_name = os.path.basename(ckpt.model_checkpoint_path)
      self.saver.restore(self.sess, os.path.join(checkpoint_dir, ckpt_name))
      counter = int(next(re.finditer("(\d+)(?!.*\d)",ckpt_name)).group(0))
      print(" [*] Success to read {}".format(ckpt_name))
      return True, counter
    else:
      print(" [*] Failed to find a checkpoint")
      return False, 0

model.py定义了DCGAN类，包括9个函数

init() 初始化参数
参数初始化，已讲过的input_height, input_width, crop, batch_size, output_height, output_width, dataset_name, input_fname_pattern, checkpoint_dir, sample_dir就不再说了
sample_numsample_num：大小和batch_size一样
y_dimy_dim：输出通道。训练mnist数据集时，y_dim=10训练mnist数据集时，y_dim=10，我想可能是因为mnist是图片数字，分为10类。如果不是mnist，则默认为none。
z_dimz_dim：噪声z的维度，默认为100
gf_dimgf_dim：G第一个卷积层的过滤器个数G第一个卷积层的过滤器个数，默认为64
df_dimdf_dim：D第一个卷积层的过滤器个数D第一个卷积层的过滤器个数，默认为64
gfc_dimgfc_dim：G第一个全连接层的G单元个数G第一个全连接层的G单元个数，默认为1024
dfc_dimdfc_dim：D第一个全连接层的D单元个数D第一个全连接层的D单元个数，默认为1024
c_dimc_dim：颜色通道，灰度图像设为1，彩色图像设为3，默认为3
其中self.d_bn1, self.d_bn2, g_bn0, g_bn1, g_bn2是batch标准化，见ops.py的batch_norm(object)。
如果是mnist数据集，d_bn3, g_bn3都要batch_norm。
self.data读取数据集。

build_model() 建立模型
inputs的形状为[batch_size, input_height, input_width, c_dim]。
如果crop=True，inputs的形状为[batch_size, output_height, output_width, c_dim]。
输入分为样本输入inputs和抽样输入sample_inputs。
噪声z的形状为[None, z_dim]，第一个None是batch的大小。
然后取数据：
self.G = self.generator(self.z)#返回[batch_size, output_height, output_width, c_dim]形状的张量,也就是batch_size张图
self.D, self.D_logits = self.discriminator(inputs)#返回的D为是否是真样本的sigmoid概率，D_logits是未经sigmoid处理
self.sampler = self.sampler(self.z)#相当于测试，经过G网络模型，取样，代码和G很像，没有G训练的过程。
self.D_, self.D_logits_ = self.discriminator(self.G, reuse=True)
#D是真实数据，D_是假数据
用交叉熵计算损失，共有：d_loss_real、d_loss_fake、g_loss
self.d_loss_real = tf.reduce_mean(
sigmoid_cross_entropy_with_logits(self.D_logits, tf.ones_like(self.D)))
self.d_loss_fake = tf.reduce_mean(
sigmoid_cross_entropy_with_logits(self.D_logits_, tf.zeros_like(self.D_)))
self.g_loss = tf.reduce_mean(
sigmoid_cross_entropy_with_logits(self.D_logits_, tf.ones_like(self.D_)))
tf.ones_like：新建一个与给定tensor大小一致的tensor，其全部元素为1
d_loss_real是真样本输入的损失，要让D_logits接近于1，也就是D识别出真样本为真的
d_loss_fake是假样本输入的损失，要让D_logits_接近于0，D识别出假样本为假
d_loss = d_loss_real + d_loss_fake是D的目标，要最小化这个损失
g_loss：要让D识别假样本为真样本，G的目标是降低这个损失，D是提高这个损失
summary这几步是关于可视化，就不管了

train()
通过Adam优化器最小化d_loss和g_loss。
sample_z为从-1到1均匀分布的数，大小为[sample_num, z_dim]
从路径中读取原始样本sample，大小为[sample_num, output_height, output_width, c_dim]
接下来进行epoch个训练：
将data总数分为batch_idxs次训练，每次训练batch_size个样本。产生的样本为batch_images。
batch_z为训练的噪声，大小为[batch_num, z_dim]
d_optim = tf.train.AdamOptimizer(config.learning_rate, beta1=config.beta1) \
.minimize(self.d_loss, var_list=self.d_vars)
g_optim = tf.train.AdamOptimizer(config.learning_rate, beta1=config.beta1) \
.minimize(self.g_loss, var_list=self.g_vars)
首先输入噪声z和batch_images，通过优化d_optim更新D网络。
然后输入噪声z，优化g_optim来更新G网络。G网络更新两次，以免d_loss为0。这点不同于paper。
这样的训练，每过100个可以生成图片看看效果。
if np.mod(counter, 100) == 1

discriminator()

代码自定义了一个conv2d，对tf.nn.conv2d稍加修改了。下面贴出tf.nn.conv2dtf.nn.conv2d解释如下：
tf.nn.conv2d(input, filter, strides, padding, use_cudnn_on_gpu=None, name=None)
除去name参数用以指定该操作的name，与方法有关的一共五个参数：
第一个参数input：指需要做卷积的输入图像，它要求是一个Tensor，具有[batch, in_height, in_width, in_channels]这样的shape，具体含义是[训练时一个batch的图片数量, 图片高度, 图片宽度, 图像通道数]，注意这是一个4维的Tensor，要求类型为float32和float64其中之一
第二个参数filter：相当于CNN中的卷积核，它要求是一个Tensor，具有[filter_height, filter_width, in_channels, out_channels]这样的shape，具体含义是[卷积核的高度，卷积核的宽度，图像通道数，卷积核个数]，要求类型与参数input相同，有一个地方需要注意，第三维in_channels，就是参数input的第四维
第三个参数strides：卷积时在图像每一维的步长，这是一个一维的向量，长度4
第四个参数padding：string类型的量，只能是”SAME”,”VALID”其中之一，这个值决定了不同的卷积方式（后面会介绍）
第五个参数：use_cudnn_on_gpu:bool类型，是否使用cudnn加速，默认为true
结果返回一个Tensor，这个输出，就是我们常说的feature map
batch_norm(object)batch_norm(object)
tf.contrib.layers.batch_norm的代码见https://github.com/tensorflow/tensorflow/blob/master/tensorflow/contrib/layers/python/layers/layers.py
batchnormalization来自于http://arxiv.org/abs/1502.03167
加快训练。

激活函数lrelu见ops.py。四次卷积（其中三次卷积之前先批标准化）和激活之后。然后线性化，返回sigmoid函数处理后的结果。h3到h4的全连接相当于线性化，用一个矩阵将h3和h4连接起来，使h4是一个batch_size维的向量。

generator()
self.h0 = tf.reshape(self.z_, [-1, s_h16, s_w16, self.gf_dim * 8])改变z_的形状。-1代表的含义是不用我们自己指定这一维的大小，函数会自动计算，但列表中只能存在一个-1。（当然如果存在多个-1，就是一个存在多解的方程了）
deconv2d()deconv2d()
引用tf的反卷积函数tf.nn.conv2d_transpose或tf.nn.deconv2d。以tf.nn.conv2d_transpose为例。
defconv2d_transpose(value, filter, output_shape, strides,padding=”SAME”, data_format=”NHWC”, name=None):

• value: 是一个4维的tensor，格式为[batch, height, width, in_channels] 或者 [batch,
  in_channels,height, width]。
• filter: 是一个4维的tensor，格式为[height, width, output_channels,
  in_channels]，过滤器的in_ channels的维度要和这个匹配。
• output_shape: 一维tensor，表示反卷积操作的输出shapeA
• strides: 针对每个输入的tensor维度，滑动窗口的步长。
• padding: “VALID”或者”SAME”，padding算法
• data_format: “NHWC”或者”NCHW” ，对应value的数据格式。
• name: 可选，返回的tensor名。

deconv= tf.nn.conv2d_transpose(input_, w, output_shape=output_shape,strides=[1,d_h, d_w, 1])
第一个参数是输入，即上一层的结果，
第二个参数是输出输出的特征图维数，是个4维的参数，
第三个参数卷积核的移动步长，[1, d_h, d_w, 1]，其中第一个对应一次跳过batch中的多少图片，第二个d_h对应一次跳过图片中多少行，第三个d_w对应一次跳过图片中多少列，第四个对应一次跳过图像的多少个通道。这里直接设置为[1，2，2，1]。即每次反卷积后，图像的滑动步长为2，特征图会扩大缩小为原来2*2=4倍。

sampler()
和generator结构一样，用的也是它的参数。存在的意义可能在于共享参数？
将self.sampler = self.sampler(self.z, self.y)改为self.sampler = self.generator(self.z, self.y)
报错：

所以sampler的存在还是有意义的。

load_mnist(), save(), load()
这三个加载保存等就不仔细讲了。

download.py和ops.py好像也没什么好讲的。
utils.py包含可视化等函数

• main.py

import os
import scipy.misc
import numpy as np

from model import DCGAN
from utils import pp, visualize, to_json, show_all_variables

import tensorflow as tf

flags = tf.app.flags
#迭代次数
flags.DEFINE_integer("epoch", 25, "Epoch to train [25]")

flags.DEFINE_float("learning_rate", 0.0002, "Learning rate of for adam [0.0002]")
flags.DEFINE_float("beta1", 0.5, "Momentum term of adam [0.5]")
flags.DEFINE_float("train_size", np.inf, "The size of train images [np.inf]")
flags.DEFINE_integer("batch_size", 64, "The size of batch images [64]")
flags.DEFINE_integer("input_height", 108, "The size of image to use (will be center cropped). [108]")
flags.DEFINE_integer("input_width", None, "The size of image to use (will be center cropped). If None, same value as input_height [None]")
flags.DEFINE_integer("output_height", 64, "The size of the output images to produce [64]")
flags.DEFINE_integer("output_width", None, "The size of the output images to produce. If None, same value as output_height [None]")
flags.DEFINE_string("dataset", "celebA", "The name of dataset [celebA, mnist, lsun]")
flags.DEFINE_string("input_fname_pattern", "*.jpg", "Glob pattern of filename of input images [*]")
flags.DEFINE_string("checkpoint_dir", "checkpoint", "Directory name to save the checkpoints [checkpoint]")
flags.DEFINE_string("sample_dir", "samples", "Directory name to save the image samples [samples]")
flags.DEFINE_boolean("train", False, "True for training, False for testing [False]")
flags.DEFINE_boolean("crop", False, "True for training, False for testing [False]")
flags.DEFINE_boolean("visualize", False, "True for visualizing, False for nothing [False]")
FLAGS = flags.FLAGS

def main(_):
  ##打印参数
  pp.pprint(flags.FLAGS.__flags)

  	
  if FLAGS.input_width is None:
    FLAGS.input_width = FLAGS.input_height
  if FLAGS.output_width is None:
    FLAGS.output_width = FLAGS.output_height

  ##检测checkpoint和sample目录是否存在
  if not os.path.exists(FLAGS.checkpoint_dir):
    os.makedirs(FLAGS.checkpoint_dir)
  if not os.path.exists(FLAGS.sample_dir):
    os.makedirs(FLAGS.sample_dir)

  ##session参数设置 tf.ConfigProto()一般用于创建session时，对参数进行设置 详见：https://blog.csdn.net/u012436149/article/details/53837651
  #gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.333)
  run_config = tf.ConfigProto()
  run_config.gpu_options.allow_growth=True

  with tf.Session(config=run_config) as sess:
     ##mnist数据集
    if FLAGS.dataset == 'mnist':
      dcgan = DCGAN(
          sess,
          input_width=FLAGS.input_width,
          input_height=FLAGS.input_height,
          output_width=FLAGS.output_width,
          output_height=FLAGS.output_height,
          batch_size=FLAGS.batch_size,
          sample_num=FLAGS.batch_size,
          y_dim=10,
          dataset_name=FLAGS.dataset,
          input_fname_pattern=FLAGS.input_fname_pattern,
          crop=FLAGS.crop,
          checkpoint_dir=FLAGS.checkpoint_dir,
          sample_dir=FLAGS.sample_dir)
    else:
      ##其他数据集
      dcgan = DCGAN(
          sess,
          input_width=FLAGS.input_width,
          input_height=FLAGS.input_height,
          output_width=FLAGS.output_width,
          output_height=FLAGS.output_height,
          batch_size=FLAGS.batch_size,
          sample_num=FLAGS.batch_size,
          dataset_name=FLAGS.dataset,
          input_fname_pattern=FLAGS.input_fname_pattern,
          crop=FLAGS.crop,
          checkpoint_dir=FLAGS.checkpoint_dir,
          sample_dir=FLAGS.sample_dir)

    show_all_variables()

    if FLAGS.train:
      dcgan.train(FLAGS)
    else:
      if not dcgan.load(FLAGS.checkpoint_dir)[0]:
        raise Exception("[!] Train a model first, then run test mode")
      

    # to_json("./web/js/layers.js", [dcgan.h0_w, dcgan.h0_b, dcgan.g_bn0],
    #                 [dcgan.h1_w, dcgan.h1_b, dcgan.g_bn1],
    #                 [dcgan.h2_w, dcgan.h2_b, dcgan.g_bn2],
    #                 [dcgan.h3_w, dcgan.h3_b, dcgan.g_bn3],
    #                 [dcgan.h4_w, dcgan.h4_b, None])

    # Below is codes for visualization
    OPTION = 2
    visualize(sess, dcgan, FLAGS, OPTION)

if __name__ == '__main__':
  tf.app.run()

main.py说明
入口程序，事先定义所需参数的值。

FLAGS参数
epochepoch：训练回合，默认为25
learning_ratelearning_rate：AdamAdam的学习率，默认为0.0002
beta1beta1：Adam的动量项（Momentum term of Adam），默认为0.5
train_sizetrain_size：训练图像的个数，默认为np.inf
batch_sizebatch_size：批图像的个数，默认为64。后面生成的图片拼在一张图，因此batch_size最好取平方，比如64，36等后面生成的图片拼在一张图，因此batch_size最好取平方，比如64，36等
input_heightinput_height：所使用的图像的图像高度（将会被center croppedcenter cropped），默认为108
input_widthinput_width：所使用的图像的图像宽度（将会被center croppedcenter cropped），如果没有特别指定默认和input_height一样
output_heightoutput_height：所产生的图像的图像高度（将会被center croppedcenter cropped），默认为64
output_widthoutput_width：所产生的图像的图像宽度（将会被center croppedcenter cropped），如果没有特别指定默认和output_height一样
datasetdataset：所用数据集的名称，在文件夹data里面，可以选择celebA，mnist，lsun。也可以自己下载图片，把文件夹放到data文件夹里面。
input_fname_patterninput_fname_pattern：输入的图片类型，默认为*.jpg
checkpoint_dircheckpoint_dir：存放checkpoint的目录名，默认为checkpointcheckpoint
sample_dirsample_dir：存放生成图片的目录名，默认为samples
traintrain：训练为True，测试为False，默认为False
cropcrop：训练为True，测试为False，默认为False
visualizevisualize：可视化为True，不可视化为False，默认为False

5. 训练
   新建data\faces目录，把头像存放到face目录下，执行

python main.py --dataset faces --input_height 96 --input_width 96 --output_height 48 --output_width 48 --crop --train --epoch 20 --input_fname_pattern "*.jpg"

由于使用笔记本的CPU训练，进程比较慢，现在大概也训练50个epoch，大概的效果：
epoch 0

epoch 20

epoch 50

epoch 100