tensorflow：使用卷积网络（CNN）实现mnist

tensorflow1.8.0+python3.6.4

#coding=utf-8
import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data
mnist=input_data.read_data_sets("MNIST_data",one_hot=True) #读取图片

def weights(shape):
    #正态分布，标准差为0.1，范围为[-1,1]
    initial=tf.truncated_normal(shape,stddev=0.1) 
    return tf.Variable(initial)

def biases(shape):
    initial=tf.constant(0.1,shape=shape)
    return tf.Variable(initial)

def conv2d(x,w):
    #strides是步长，strides[0]=strides[3]=1(这是规定的），strides[1]表示横向步长，strides[2]表示纵向步长
    #SAME表示边缘外自动补0
    return tf.nn.conv2d(x,w,strides=[1,1,1,1],padding='SAME') 

def max_pool_2x2(x):
    #ksize是池化形状大小，ksize[0]=ksize[3]=1(这是规定的），ksize[1]表示横向的长度，ksize[2]表示纵向大小
    #strides是步长，strides[0]=strides[3]=1(这是规定的），strides[1]表示横向步长，strides[2]表示纵向步长
    #SAME表示边缘外自动补0
    return tf.nn.max_pool(x,ksize=[1,2,2,1],strides=[1,2,2,1],padding='SAME')

xs=tf.placeholder(tf.float32,[None,28*28])
ys=tf.placeholder(tf.float32,[None,10])
x_image=tf.reshape(xs,[-1,28,28,1])

#The first conv_layer
w_conv1=weights([5,5,1,32])
b_conv1=biases([32])
h_conv1=tf.nn.relu(conv2d(x_image,w_conv1)+b_conv1)
h_pool1= max_pool_2x2(h_conv1)

#The second conv_layer
w_conv2=weights([5,5,32,64])
b_conv2=biases([64])
h_conv2=tf.nn.relu(conv2d(h_pool1,w_conv2)+b_conv2)
h_pool2=max_pool_2x2(h_conv2)

#The third layer
w_fc1=weights([7*7*64,1024])
b_fc1=biases([1024])
h_pool2_flat=tf.reshape(h_pool2,[-1,7*7*64])
h_fc1=tf.nn.relu(tf.matmul(h_pool2_flat,w_fc1)+b_fc1)

#overfiting
keep_prob=tf.placeholder(tf.float32)
h_fc1_drop=tf.nn.dropout(h_fc1,keep_prob)

#The fourth layer
w_fc2=weights([1024,10])
b_fc2=biases([10])
y_conv=tf.nn.softmax(tf.matmul(h_fc1_drop,w_fc2)+b_fc2)

cross_entropy=-tf.reduce_sum(ys*tf.log(y_conv))
train_step=tf.train.GradientDescentOptimizer(0.5).minimize(cross_entropy)

correct_prediction=tf.equal(tf.argmax(y_conv,1),tf.argmax(ys,1))
accuracy=tf.reduce_mean(tf.cast(correct_prediction,tf.float32))
init=tf.global_variables_initializer()
sess=tf.Session()
sess.run(init)

for i in range(1000):
    batch=mnist.train.next_batch(50)
    #if i%100==0:
    #    train_accuracy=accuracy.eval(session=sess,feed_dict={xs:batch[0],ys:batch[1],keep_prob:1.0})
    #    #print("step %d, training accurary %g"%(i,train_accuracy))
    train_step.run(session=sess,feed_dict={xs:batch[0],ys:batch[1],keep_prob:1.0})
print("test accurary %g"%accuracy.eval(session=sess,feed_dict={xs:mnist.test.images,ys:mnist.test.labels,keep_prob:1.0}))