一、前述
GAN，生成对抗网络，在2016年基本火爆深度学习，所有有必要学习一下。生成对抗网络直观的应用可以帮我们生成数据，图片。
二、具体
1、生活案例
比如假设真钱 r
坏人定义为G 我们通过 G 给定一个噪音X 通过学习一组参数w 生成一个G（x），转换成一个真实的分布。 这就是生成，相当于造假钱。
警察定义为D 将G(x)和真钱r 分别输入给判别网络，能判别出真假，真钱判别为0，假钱判别为1 。这就是判别。
最后生成网络想让判别网络判别不出来什么是真实的，什么是假的。要想生成的更好，则判别的就必须更强。有些博弈的思想，只有你强了，我才更强！！。
2、数学案例
我们最后的希望。
3、损失函数
4、代码案例
流程：
为了使判别模型更好，所以我们额外训练一个D_pre网络，使得判别模型能够判别出哪些是0，哪些是1，训练完之后会得到一组w,b参数。这样我们在真正初始化判别模型D的时候就能根据之前的D_pre来进行初始化。
代码：
import argparse
import numpy as np
from scipy.stats import norm
import tensorflow as tf
import matplotlib.pyplot as plt
from matplotlib import animation
import seaborn as sns sns.set(color_codes=True) seed = 42
np.random.seed(seed)
tf.set_random_seed(seed) class DataDistribution(object):
def __init__(self):
self.mu = 4#均值
self.sigma = 0.5#标准差 def sample(self, N):
samples = np.random.normal(self.mu, self.sigma, N)
samples.sort()
return samples class GeneratorDistribution(object):#在生成模型额噪音点，初始化输入
def __init__(self, range):
self.range = range def sample(self, N):
return np.linspace(-self.range, self.range, N) + \
np.random.random(N) * 0.01 def linear(input, output_dim, scope=None, stddev=1.0):
norm = tf.random_normal_initializer(stddev=stddev)
const = tf.constant_initializer(0.0)
with tf.variable_scope(scope or 'linear'):
w = tf.get_variable('w', [input.get_shape()[1], output_dim], initializer=norm)
b = tf.get_variable('b', [output_dim], initializer=const)
return tf.matmul(input, w) + b def generator(input, h_dim):
h0 = tf.nn.softplus(linear(input, h_dim, 'g0'))#12*1
h1 = linear(h0, 1, 'g1')
return h1#z最后的生成模型 def discriminator(input, h_dim):
h0 = tf.tanh(linear(input, h_dim * 2, 'd0'))#linear 控制初始化参数
h1 = tf.tanh(linear(h0, h_dim * 2, 'd1'))
h2 = tf.tanh(linear(h1, h_dim * 2, scope='d2')) h3 = tf.sigmoid(linear(h2, 1, scope='d3'))#最终的输出值 对判别网络输出
return h3 def optimizer(loss, var_list, initial_learning_rate):
decay = 0.95
num_decay_steps = 150#没迭代150次 学习率衰减一次0.95-150*0.95
batch = tf.Variable(0)
learning_rate = tf.train.exponential_decay(
initial_learning_rate,
batch,
num_decay_steps,
decay,
staircase=True
)
optimizer = tf.train.GradientDescentOptimizer(learning_rate).minimize(
loss,
global_step=batch,
var_list=var_list
)
return optimizer class GAN(object):
def __init__(self, data, gen, num_steps, batch_size, log_every):
self.data = data
self.gen = gen
self.num_steps = num_steps
self.batch_size = batch_size
self.log_every = log_every
self.mlp_hidden_size = 4#隐层神经元个数 self.learning_rate = 0.03#学习率 self._create_model() def _create_model(self): with tf.variable_scope('D_pre'):#构造D_pre模型骨架,预先训练,为了去初始化真正的判别模型
self.pre_input = tf.placeholder(tf.float32, shape=(self.batch_size, 1))
self.pre_labels = tf.placeholder(tf.float32, shape=(self.batch_size, 1))
D_pre = discriminator(self.pre_input, self.mlp_hidden_size)
self.pre_loss = tf.reduce_mean(tf.square(D_pre - self.pre_labels))
self.pre_opt = optimizer(self.pre_loss, None, self.learning_rate) # This defines the generator network - it takes samples from a noise
# distribution as input, and passes them through an MLP.
with tf.variable_scope('Gen'):#生成模型
self.z = tf.placeholder(tf.float32, shape=(self.batch_size, 1))#噪音的输入
self.G = generator(self.z, self.mlp_hidden_size)#最后的生成结果 # The discriminator tries to tell the difference between samples from the
# true data distribution (self.x) and the generated samples (self.z).
#
# Here we create two copies of the discriminator network (that share parameters),
# as you cannot use the same network with different inputs in TensorFlow.
with tf.variable_scope('Disc') as scope:#判别模型 不光接受真实的数据 还要接受生成模型的判别
self.x = tf.placeholder(tf.float32, shape=(self.batch_size, 1))
self.D1 = discriminator(self.x, self.mlp_hidden_size)#真实的数据
scope.reuse_variables()#变量重用
self.D2 = discriminator(self.G, self.mlp_hidden_size)#生成的数据 # Define the loss for discriminator and generator networks (see the original
# paper for details), and create optimizers for both
self.loss_d = tf.reduce_mean(-tf.log(self.D1) - tf.log(1 - self.D2))#判别网络的损失函数
self.loss_g = tf.reduce_mean(-tf.log(self.D2))#生成网络的损失函数,希望其趋向于1 self.d_pre_params = tf.get_collection(tf.GraphKeys.TRAINABLE_VARIABLES, scope='D_pre')
self.d_params = tf.get_collection(tf.GraphKeys.TRAINABLE_VARIABLES, scope='Disc')
self.g_params = tf.get_collection(tf.GraphKeys.TRAINABLE_VARIABLES, scope='Gen') self.opt_d = optimizer(self.loss_d, self.d_params, self.learning_rate)
self.opt_g = optimizer(self.loss_g, self.g_params, self.learning_rate) def train(self):
with tf.Session() as session:
tf.global_variables_initializer().run() # pretraining discriminator
num_pretrain_steps = 1000#迭代次数，先训练D_pre ,先让其有一个比较好的初始化参数
for step in range(num_pretrain_steps):
d = (np.random.random(self.batch_size) - 0.5) * 10.0
labels = norm.pdf(d, loc=self.data.mu, scale=self.data.sigma)
pretrain_loss, _ = session.run([self.pre_loss, self.pre_opt], {#相当于一次迭代
self.pre_input: np.reshape(d, (self.batch_size, 1)),
self.pre_labels: np.reshape(labels, (self.batch_size, 1))
})
self.weightsD = session.run(self.d_pre_params)#相当于拿到之前的参数
# copy weights from pre-training over to new D network
for i, v in enumerate(self.d_params):
session.run(v.assign(self.weightsD[i]))#吧权重参数拷贝 for step in range(self.num_steps):#训练真正的生成对抗网络
# update discriminator
x = self.data.sample(self.batch_size)#真实的数据
z = self.gen.sample(self.batch_size)#随意的数据，噪音点
loss_d, _ = session.run([self.loss_d, self.opt_d], {#D两种输入真实，和生成的
self.x: np.reshape(x, (self.batch_size, 1)),
self.z: np.reshape(z, (self.batch_size, 1))
}) # update generator
z = self.gen.sample(self.batch_size)#G网络
loss_g, _ = session.run([self.loss_g, self.opt_g], {
self.z: np.reshape(z, (self.batch_size, 1))
}) if step % self.log_every == 0:
print('{}: {}\t{}'.format(step, loss_d, loss_g))
if step % 100 == 0 or step==0 or step == self.num_steps -1 :
self._plot_distributions(session) def _samples(self, session, num_points=10000, num_bins=100):
xs = np.linspace(-self.gen.range, self.gen.range, num_points)
bins = np.linspace(-self.gen.range, self.gen.range, num_bins) # data distribution
d = self.data.sample(num_points)
pd, _ = np.histogram(d, bins=bins, density=True) # generated samples
zs = np.linspace(-self.gen.range, self.gen.range, num_points)
g = np.zeros((num_points, 1))
for i in range(num_points // self.batch_size):
g[self.batch_size * i:self.batch_size * (i + 1)] = session.run(self.G, {
self.z: np.reshape(
zs[self.batch_size * i:self.batch_size * (i + 1)],
(self.batch_size, 1)
)
})
pg, _ = np.histogram(g, bins=bins, density=True) return pd, pg def _plot_distributions(self, session):
pd, pg = self._samples(session)
p_x = np.linspace(-self.gen.range, self.gen.range, len(pd))
f, ax = plt.subplots(1)
ax.set_ylim(0, 1)
plt.plot(p_x, pd, label='real data')
plt.plot(p_x, pg, label='generated data')
plt.title('1D Generative Adversarial Network')
plt.xlabel('Data values')
plt.ylabel('Probability density')
plt.legend()
plt.show()
def main(args):
model = GAN(
DataDistribution(),
GeneratorDistribution(range=8),
args.num_steps,
args.batch_size,
args.log_every,
)
model.train() def parse_args():
parser = argparse.ArgumentParser()
parser.add_argument('--num-steps', type=int, default=1200,
help='the number of training steps to take')
parser.add_argument('--batch-size', type=int, default=12,
help='the batch size')
parser.add_argument('--log-every', type=int, default=10,
help='print loss after this many steps')
return parser.parse_args() if __name__ == '__main__':
main(parse_args())
结果：
迭代到最后时候可以看到结果越来越类似。