基于Tensorflow+Python实现降噪自动编码器(DAE)

It is our choices that show what we truly are, far more than our abilities.

决定我们一生的,不是我们的能力,而是我们的选择。

基于Tensorflow+Python实现降噪自动编码器(DAE)

本代码原理参考书籍TensorFlow 实战

# -*- coding: UTF-8 -*-

'''
Created on 2017年8月7日

@summary:利用tensorflow实现降噪自动编码器

@author: dreamhome
'''

import numpy as np
import sklearn.preprocessing as prep
import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data


def xavier_init(n_input,n_output,constant=1):
    """
    Xavier初始化器 让权重被初始化调整合理的分布 mean=0 std=2/(n_input+n_output)
    :param n_input:输入节点数量
    :param n_output:输出节点数量
    """
    low=-constant * np.sqrt(6.0/(n_input+n_output))
    high=constant * np.sqrt(6.0/(n_input+n_output))
    return tf.random_uniform((n_input,n_output), minval=low, maxval=high, 
                             dtype=tf.float32)



class AdditiveGaussianNoiseAutoencoder(object):
    def __init__(self,n_input,n_hidden,transfer_function=tf.nn.softplus,
                 optimizer=tf.train.AdamOptimizer(),scale=0.1):
        """
        初始化函数
        :param n_input:输入变量数
        :param n_hidden:隐含层节点数
        :param transfer_function:隐含层激活函数
        :param optimizer:优化器默认为 Adam
        :param scale:高斯噪声系数
        """
        self.n_input=n_input
        self.n_hidden=n_hidden
        self.transfer=transfer_function
        self.scale=tf.placeholder(dtype=tf.float32)
        self.training_scale=scale

        network_weights=self._initialize_weights()
        self.weights=network_weights

        self.x=tf.placeholder(tf.float32,[None,self.n_input])
        #利用transform对结果进行激活函数处理
        self.hidden=self.transfer(tf.add(tf.matmul(
            self.x+scale*tf.random_normal((n_input,)),self.weights['w1'])
                                          ,self.weights['b1']))
        #reconstruction层对经过隐含层的数据进行复原
        self.reconstruction=tf.add(tf.matmul(
            self.hidden,self.weights['w2']),
                                   self.weights['b2'])

        #定义自编码器的损失函数 平方误差
        self.cost=0.5 * tf.reduce_sum(tf.pow(tf.subtract(self.reconstruction, self.x),2.0))
        self.optimizer=optimizer.minimize(self.cost)

        init=tf.global_variables_initializer()
        self.sess=tf.Session()
        self.sess.run(init)



    def _initialize_weights(self):
        """
        参数初始化函数
        """
        #利用字典存储参数
        all_weights=dict()

        all_weights['w1']=tf.Variable(xavier_init(self.n_input,self.n_hidden))
        all_weights['b1']=tf.Variable(tf.zeros([self.n_hidden], dtype=tf.float32))
        all_weights['w2']=tf.Variable(tf.zeros([self.n_hidden,self.n_input], dtype=tf.float32))
        all_weights['b2']=tf.Variable(tf.zeros([self.n_input], dtype=tf.float32))

        return all_weights


    def partial_fit(self,X):
        """
        函数利用一个batch的数据进行训练并且返回当前的损失cost
        :param X: 输入数据X
        """
        cost,opt=self.sess.run((self.cost,self.optimizer),
                           feed_dict={self.x:X,
                                      self.scale:self.training_scale})
        return cost


    def calc_total_cost(self,X):
        """
        函数在测试集上对模型进行评测时用到
        """
        return self.sess.run(self.cost,
                         feed_dict={self.x:X,
                                    self.scale:self.training_scale})

    def transform(self,X):
        """
        返回自编码器隐含层的结果,提供一个接口来获取抽象后的特征
        """
        return self.sess.run(self.hidden,feed_dict={self.x:X,
                                                    self.scale:self.training_scale})

    def generate(self,hidden=None):
        """
        将高阶特征复原为原始数据
        """
        if hidden is None:
            hidden=np.random.normal(size=self.weights["b1"])
        return self.sess.run(self.reconstruction,
                             feed_dict={self.hidden:hidden})

    def reconstruct(self,X):
        """
        函数整体运行一遍复原过程 包括高阶特征提取和通过高阶特征复原数据
        """
        return self.sess.run(self.reconstruct,feed_dict={self.x:X,
                                                         self.scale:self.training_scale})

    def getWeights(self):
        """
        获取隐含层的权重w1
        """
        return self.sess.run(self.weights['w1'])

    def getBiases(self):
        """
        获取隐含层偏置系数b1
        """
        return self.sess.run(self.weights['b1'])

def standard_scale(X_train,X_test):
    """
    对训练集和测试集的数据进行标准化处理
    """
    preprocessor =prep.StandardScaler().fit(X_train)
    X_train=preprocessor.transform(X_train)
    X_test=preprocessor.transform(X_test)
    return X_train,X_test

def get_random_block_from_data(data,batch_size):
    """
    函数随机从数据集中获取block
    """
    start_index=np.random.randint(0,len(data)-batch_size)
    return data[start_index:(start_index+batch_size)]


if __name__ == '__main__':

    mnist=input_data.read_data_sets('MNIST_data',one_hot=True)
    X_train,X_test=standard_scale(mnist.train.images, mnist.test.images)
    n_samples=int(mnist.train.num_examples)
    training_epochs=1
    batch_size=128
    display_step=1

    autoencoder=AdditiveGaussianNoiseAutoencoder(n_input=784,
                                                 n_hidden=200,
                                                 transfer_function=tf.nn.softplus,
                                                 optimizer=tf.train.AdamOptimizer(learning_rate=0.001),
                                                 scale=0.01)

    for epoch in range(training_epochs):
        avg_cost=0.0
        total_batch=int(n_samples/batch_size)
        for i in range(total_batch):
            batch_xs=get_random_block_from_data(X_train, batch_size)
            cost=autoencoder.partial_fit(batch_xs)
            avg_cost+=cost/n_samples*batch_size
        if epoch%display_step==0:
            print("Epoch:",'%04d'%(epoch+1),
                  "Cost=","{:.9f}".format(avg_cost))
    print("Total cost:"+str(autoencoder.calc_total_cost(X_test)))

运行结果截图如下:
基于Tensorflow+Python实现降噪自动编码器(DAE)_第1张图片

以上是本人对于此内容的理解,敬请广大读者随时不吝批评指正,感谢。

你可能感兴趣的:(deep,learning)