TensorFlow的学习之路--MNIST算法解析

# -*- coding: utf-8 -*-
"""
Created on Thu Mar  8 16:28:13 2018

@author: kxq
"""
import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data
mnist=input_data.read_data_sets('MNIST_data',one_hot=True)

def compute_accuracy(v_xs,v_ys):
    global prediction
    y_pre=sess.run(prediction,feed_dict={xs:v_xs,keep_prob:1})
    correct_prediction=tf.equal(tf.argmax(y_pre,1),tf.argmax(v_ys,1))
    accuracy=tf.reduce_mean(tf.cast(correct_prediction,tf.float32))
    result=sess.run(accuracy,feed_dict={xs:v_xs,ys:v_ys,keep_prob:1})
    return result

def weight_variable(shape):
    inital=tf.truncated_normal(shape,stddev=0.1)##产生随机变量，与np.random.normal一样的功能
    return tf.Variable(inital)
def bias_variable(shape):
    inital=tf.constant(0.1,shape=shape)
    return tf.Variable(inital)
def conv2d(x,W):
    ## 格式为[1,x_movement,y_movement,1] 步长为1*1
    ## 最外面与原图形一直 
    ## 卷积层需要传入W
    return tf.nn.conv2d(x,W,strides=[1,1,1,1],padding='SAME')
def max_pool_2x2(x):
    ## pool层不需要传入W
    ##格式为[1,x_movement,y_movement,1] 步长为2*2
    return tf.nn.max_pool(x,ksize=[1,2,2,1],strides=[1,2,2,1],padding='SAME')

#定义输入层的占有符
xs=tf.placeholder(shape=[None,784],dtype=tf.float32) #28*28
ys=tf.placeholder(shape=[None,10],dtype=tf.float32)
keep_prob=tf.placeholder(tf.float32)
##因为xs传入的是所有数据，-1代表是不确定纬度，现在要转成28*28,1代表是灰度图像
x_image=tf.reshape(xs,[-1,28,28,1])
##print(x_image.shape) ##显示[n_samples,28,28,1]

##conv1 layer1##
##5*5的patch进行扫描
##1是输入的图像
##高度为32
W_conv1=weight_variable([5,5,1,32])##patch size 5*5 ,输入 size 1 ,outsize:32
b_conv1=bias_variable([32])
##嵌套tf.nn.relu是让卷积进行非线性化激活
h_conv1=tf.nn.relu(conv2d(x_image,W_conv1)+b_conv1)##output size=28*28*32
h_pool1=max_pool_2x2(h_conv1)                      ##output size 14*14*32


##conv2 layer2##
W_conv2=weight_variable([5,5,32,64])##还是以5*5的patch，但输入size 为32 ，设定输出size 64
b_conv2=bias_variable([64])
h_conv2=tf.nn.relu(conv2d(h_pool1,W_conv2)+b_conv2) #output size =14*14*64
h_pool2=max_pool_2x2(h_conv2)                       #output size =7*7*64

##func1 layer1##
W_fc1=weight_variable([7*7*64,1024]) ##定义1024是让输出更大，这个值可以改变
b_fc1=bias_variable([1024])
##[n_sample,7,7,64]->>[n_sample,7*7*64]
h_pool_flat=tf.reshape(h_pool2,[-1,7*7*64])
h_fc1=tf.nn.relu(tf.matmul(h_pool_flat,W_fc1)+b_fc1) 
##dropout处理
##h_fc1_drop=tf.nn.dropout(h_fc1,keep_prob)


##func2 layer2##
W_fc2=weight_variable([1024,10]) ##传入是1024，传出是10，因为最终结果就10个数字
b_fc2=bias_variable([10])
prediction=tf.nn.softmax(tf.matmul(h_fc1,W_fc2)+b_fc2)
#用softmax配合cross_entropy生成分类算法，计算loss
cross_entropy=tf.reduce_mean(-tf.reduce_sum(ys*tf.log(prediction),reduction_indices=[1]))#计算loss
#train
train_step=tf.train.AdamOptimizer(0.0001).minimize(cross_entropy)  
#初始化
sess=tf.Session()
sess.run(tf.initialize_all_variables())
#开始训练
for i in range(1000):
     batch_xs,batch_ys=mnist.train.next_batch(100)
     sess.run(train_step,feed_dict={xs:batch_xs,ys:batch_ys})
     if i%50==0:
         print(compute_accuracy(mnist.test.images,mnist.test.labels))