卷积神经网络应用于tensorflow手写数字识别（第三版）

import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data

mnist = input_data.read_data_sets("F:\TensorflowProject\MNIST_data",one_hot=True)

#每个批次大小
batch_size = 100
#计算一共有多少个批次
n_batch = mnist.train.num_examples //batch_size

#初始化权值
def weight_variable(shape):
initial = tf.truncated_normal(shape,stddev=0.1) #初始化一个截断的正态分布
return tf.Variable(initial)

#初始化偏值
def bias_variable(shape):
initial = tf.constant(0.1,shape=shape)
return tf.Variable(initial)

#卷积层
def conv2d(x,W):
#x input tensor of shape '[batch,in_height,in_width,in_channels]'
#W filter/kernel tensor of shape [filter_height,filter_width,in_channels,out_channels]
#strides[0] = strides[3] = 1, strides[1]代表x方向的步长，strides[2]代表y方向的步长
#padding:A string from :SAME 或者 VALID
return tf.nn.conv2d(x,W,strides=[1,1,1,1],padding='SAME')

#池化层
def max_pool_2x2(x):
#ksize[1,x,y,1]
return tf.nn.max_pool(x,ksize=[1,2,2,1],strides=[1,2,2,1],padding='SAME')

#定义两个placeholder
x = tf.placeholder(tf.float32,[None,784]) #28*28
y = tf.placeholder(tf.float32,[None,10])

#设置x的格式为4D向量 [batch,in_height,in_width,in_chanels]
x_image = tf.reshape(x,[-1,28,28,1])

#初始化第一个卷积层的权值和偏值
W_conv1 = weight_variable([5,5,1,32])
b_conv1 = bias_variable([32])

#把x_image和权值向量进行卷积，再加上偏置值，然后应用于relu激活函数
h_conv1 = tf.nn.relu(conv2d(x_image,W_conv1) + b_conv1)
h_pool1 = max_pool_2x2(h_conv1) #max-pooling，经过池化计算得到一个结果

#初始化第二个卷积层的权值和偏置值
W_conv2 = weight_variable([5,5,32,64])
b_conv2 = bias_variable([64])

#把h_pool1和权值向量进行卷积，再加上偏置值，然后应用于relu激活函数
h_conv2 = tf.nn.relu(conv2d(h_pool1,W_conv2) + b_conv2)
h_pool2 = max_pool_2x2(h_conv2) #max-pooling

#28*28的图片第一次卷积后还是28*28，第一次池化后为14*14
#第二次卷积后是14*14，第二次池化后为7*7
#上面步骤完成以后得到64张7*7的平面

#初始化第一个全连接层的权值
W_fc1 = weight_variable([7*7*64,1024]) #上一步有 7*7*64个神经元，全连接层有1024个神经元
b_fc1 = bias_variable([1024]) #1024个节点

#把池化层2的输出扁平化为1维
h_pool2_flat = tf.reshape(h_pool2,[-1,7*7*64])
#求第一个全连接层的输出
h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat,W_fc1) + b_fc1)

#keep_prob标识神经元输出概率
keep_prob =tf.placeholder(tf.float32)
h_fc1_drop = tf.nn.dropout(h_fc1,keep_prob)

#初始化第二个全连接层
W_fc2 = weight_variable([1024,10])
b_fc2 = bias_variable([10])

#计算输出
prediction = tf.nn.softmax(tf.matmul(h_fc1_drop,W_fc2) + b_fc2)

#交叉熵代价函数
cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=y,logits=prediction))
#使用AdamOptimizer进行优化
train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)

#用布尔列表存放结果
correct_prediction = tf.equal(tf.argmax(prediction,1),tf.argmax(y,1))
#求准确率
accuracy = tf.reduce_mean(tf.cast(correct_prediction,tf.float32))

with tf.Session() as sess:
　　sess.run(tf.global_variables_initializer())
　　for epoch in range(21):
　　　　for batch in range(n_batch):
　　　　　　batch_xs,batch_ys = mnist.train.next_batch(batch_size)
　　　　　　sess.run(train_step,feed_dict={x:batch_xs,y:batch_ys,keep_prob:0.7})

　　　　test_acc = sess.run(accuracy,feed_dict={x:mnist.test.images,y:mnist.test.labels,keep_prob:1.0})
　　　　print("Iter "+str(epoch)+" ,Testing Accuracy = "+str(test_acc))

##############运行结果

Iter  0  ,Testing Accuracy =  0.9552
Iter  1  ,Testing Accuracy =  0.9743
Iter  2  ,Testing Accuracy =  0.9796
Iter  3  ,Testing Accuracy =  0.9807
Iter  4  ,Testing Accuracy =  0.9849
Iter  5  ,Testing Accuracy =  0.9863
Iter  6  ,Testing Accuracy =  0.9859
Iter  7  ,Testing Accuracy =  0.9885
Iter  8  ,Testing Accuracy =  0.9887
Iter  9  ,Testing Accuracy =  0.9894
Iter  10  ,Testing Accuracy =  0.9907
Iter  11  ,Testing Accuracy =  0.991
Iter  12  ,Testing Accuracy =  0.9903
Iter  13  ,Testing Accuracy =  0.992
Iter  14  ,Testing Accuracy =  0.9904
Iter  15  ,Testing Accuracy =  0.9915
Iter  16  ,Testing Accuracy =  0.9903
Iter  17  ,Testing Accuracy =  0.9912
Iter  18  ,Testing Accuracy =  0.9917
Iter  19  ,Testing Accuracy =  0.9912
Iter  20  ,Testing Accuracy =  0.992