『TensorFlow』读书笔记_降噪自编码器

『TensorFlow』降噪自编码器设计

 之前学习过的代码，又敲了一遍，新的收获也还是有的，因为这次注释写的比较详尽，所以再次记录一下，具体的相关知识查阅之前写的文章即可（见上面链接）。

# Author : Hellcat
# Time   : 2017/12/6

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

def xavier_init(fan_in,fan_out, constant = 1):
    '''
    xavier 权重初始化方式
    :param fan_in: 行数
    :param fan_out: 列数
    :param constant: 常数权重，调节初始化范围的倍数
    :return: 初始化后的权重tensor
    '''
    low = -constant * np.sqrt(6.0 / (fan_in + fan_out))
    high = constant * np.sqrt(6.0 / (fan_in + fan_out))
    return tf.random_uniform((fan_in, fan_out),
                             minval=low, maxval=high)

class AdditiveGaussianNoiseAutoencoder():

    def __init__(self, n_input, n_hidden,
                 transfer_function=tf.nn.softplus,
                 optimizer=tf.train.AdamOptimizer(),scale=0.1):
        '''
        初始化自编码器
        :param n_input: 输入层结点数
        :param n_hidden: 隐藏层节点数
        :param transfer_function: 隐藏层激活函数
        :param optimizer: 优化器，是实例化的对象
        :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(tf.float32) # 实际网络中调用的
        self.training_scale = scale # 训练用噪声系数
        network_weights = self._initialize_weights()
        self.weights = network_weights

        self.x = tf.placeholder(tf.float32, [None, self.n_input])
        self.hidden = \
            self.transfer(
                tf.add(
                    tf.matmul(
                        self.x + self.scale * tf.random_normal((n_input,)),
                        self.weights['w1']),
                    self.weights['b1']))

        # 重建部分没有使用激活函数
        self.reconstruction = \
            tf.add(
                tf.matmul(
                    self.hidden, self.weights['w2']),
                self.weights['b2'])

        self.cost = 0.5 * tf.reduce_sum(tf.pow(tf.subtract(self.reconstruction,self.x),2.0))
        # 可以将类的实例过程作为实参传入函数
        self.optimizer = optimizer.minimize(self.cost)

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

    def _initialize_weights(self):
        '''
        初始化全部变量
        :return: 装有变量的字典
        '''
        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):
        '''
        进行单次训练并返回loss
        :param X: 训练数据
        :return: 本次损失函数值
        '''
        cost, opt = self.sess.run((self.cost, self.optimizer),
                                  feed_dict={self.x:X, self.scale:self.training_scale})
        return cost

    def calc_totul_cost(self, X):
        '''
        计算损失函数，不触发训练
        :param X: 训练数据
        :return: 损失函数
        '''
        return self.sess.run(self.cost, feed_dict={self.x:X, self.scale:self.training_scale})

    def transform(self, X):
        '''
        返回隐藏层输出结果，目的是获取抽象后的特征
        :param X: 训练数据
        :return: 隐藏层输出
        '''
        return self.sess.run(self.hidden, feed_dict={self.x:X, self.scale:self.training_scale})

    def generate(self, hidden=None):
        '''
        通过隐藏层特征重建
        :param hidden: 隐藏层特征
        :return: 重建数据
        '''
        if hidden is None:
            hidden = np.random.normal(size=[self.n_input])
        return self.sess.run(self.reconstruction, feed_dict={self.hidden:hidden})

    def reconstruct(self,X):
        '''
        从原始数据重建
        :param X: 训练数据
        :return: 重建数据
        '''
        return self.sess.run(self.reconstruction,
                             feed_dict={self.x:X, self.scale:self.training_scale})

    def getWeights(self):
        '''
        获取参数值
        :return: 隐藏层权重
        '''
        return self.sess.run(self.weights['w1'])

    def getBaises(self):
        '''
        获取参数值
        :return: 隐藏层偏置
        '''
        return self.sess.run(self.weights['b1'])

def standard_scale(X_train, X_test):
    '''
    标准化数据
    :param X_train: 训练数据
    :param X_test: 测试数据
    :return: 标准化之后的训练、测试数据
    '''
    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):
    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)
    train_epochs = 20
    batch_size = 20
    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(train_epochs):
        avg_cost = 0.
        totu_batch = int(n_samples / batch_size)
        for i in range(totu_batch):
            batch_xs = get_random_block_from_data(X_train, batch_size)

            # 单数据块训练并计算损失函数
            cost = autoencoder.partial_fit(batch_xs)
            avg_cost += cost / n_samples * batch_size

            if epoch % display_step == 0:
                print('epoch : %04d, cost = %.9f' % (epoch + 1,avg_cost))

            # 计算测试集上的cost
    print('Total coat:',str(autoencoder.calc_totul_cost(X_test)))

部分输出如下：

……

epoch : 0020, cost = 1509.876800515
epoch : 0020, cost = 1510.107261985
epoch : 0020, cost = 1510.332509055
epoch : 0020, cost = 1510.551538707
Total coat: 768927.0

1.xavier初始化权重方法

2.函数实参可以是class()，即实例化的类