原理就不多讲了，直接上代码，有详细注释。
#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)
#每个批次的大小
batch_size = 100
n_batch = mnist.train._num_examples // batch_size
def weight_variable(shape):
initial = tf.truncated_normal(shape,stddev=0.1) #生成一个截断的正态分布
return tf.Variable(initial)
def bias_variable(shape):
initial = tf.constant(0.1,shape = shape)
return tf.Variable(initial)
#卷基层
def conv2d(x,W):
return tf.nn.conv2d(x,W,strides=[1,1,1,1],padding='SAME')
#池化层
def max_pool_2x2(x):
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])
y = tf.placeholder(tf.float32,[None,10])
#改变x的格式转为4D的向量[batch,in_height,in_width,in_channels]
x_image = tf.reshape(x, [-1,28,28,1])
#初始化第一个卷基层的权值和偏置
W_conv1 = weight_variable([5,5,1,32]) #5*5的采样窗口 32个卷积核从一个平面抽取特征 32个卷积核是自定义的
b_conv1 = bias_variable([32]) #每个卷积核一个偏置值
#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) #进行max-pooling
#初始化第二个卷基层的权值和偏置
W_conv2 = weight_variable([5,5,32,64]) # 5*5的采样窗口 64个卷积核从32个平面抽取特征 由于前一层操作得到了32个特征图
b_conv2 = bias_variable([64]) #每一个卷积核一个偏置值
#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) #进行max-pooling
#28x28的图片第一次卷积后还是28x28 第一次池化后变为14x14
#第二次卷积后 变为14x14 第二次池化后变为7x7
#通过上面操作后得到64张7x7的平面
#初始化第一个全连接层的权值
W_fc1 = weight_variable([7*7*64,1024])#上一层有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
#把第二个池化层的输出扁平化为一维
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用来表示神经元的输出概率
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)
#结果存放在一个布尔列表中
correct_prediction = tf.equal(tf.argmax(prediction,1),tf.argmax(y,1)) #argmax返回一维张量中最大的值所在的位置
#求准确率
accuracy = tf.reduce_mean(tf.cast(correct_prediction,tf.float32))
saver = tf.train.Saver()
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
for epoch in range(50):
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})
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(acc))
saver.save(sess,save_path='/home/xxx/logs/mnistmodel',global_step=1)#将训练出来的权重参数保存
#测试过程肯定会出先资源超出范围因为测试的样本太大了，进行下面修改
#coding:utf-8
import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data
import os
os.environ['CUDA_VISIBLE_DEVICES']='0'
mnist = input_data.read_data_sets('MNIST_data',one_hot=True)
#每个批次的大小
batch_size = 100
n_batch = mnist.train._num_examples // batch_size
def weight_variable(shape):
initial = tf.truncated_normal(shape,stddev=0.1) #生成一个截断的正态分布
return tf.Variable(initial)
def bias_variable(shape):
initial = tf.constant(0.1,shape = shape)
return tf.Variable(initial)
#卷基层
def conv2d(x,W):
return tf.nn.conv2d(x,W,strides=[1,1,1,1],padding='SAME')
#池化层
def max_pool_2x2(x):
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])
y = tf.placeholder(tf.float32,[None,10])
#改变x的格式转为4D的向量[batch,in_height,in_width,in_channels]
x_image = tf.reshape(x, [-1,28,28,1])
#初始化第一个卷基层的权值和偏置
W_conv1 = weight_variable([5,5,1,32]) #5*5的采样窗口 32个卷积核从一个平面抽取特征 32个卷积核是自定义的
b_conv1 = bias_variable([32]) #每个卷积核一个偏置值
#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) #进行max-pooling
#初始化第二个卷基层的权值和偏置
W_conv2 = weight_variable([5,5,32,64]) # 5*5的采样窗口 64个卷积核从32个平面抽取特征 由于前一层操作得到了32个特征图
b_conv2 = bias_variable([64]) #每一个卷积核一个偏置值
#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) #进行max-pooling
#28x28的图片第一次卷积后还是28x28 第一次池化后变为14x14
#第二次卷积后 变为14x14 第二次池化后变为7x7
#通过上面操作后得到64张7x7的平面
#初始化第一个全连接层的权值
W_fc1 = weight_variable([7*7*64,1024])#上一层有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
#把第二个池化层的输出扁平化为一维
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用来表示神经元的输出概率
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)
#结果存放在一个布尔列表中
correct_prediction = tf.equal(tf.argmax(prediction,1),tf.argmax(y,1)) #argmax返回一维张量中最大的值所在的位置
#求准确率
accuracy = tf.reduce_mean(tf.cast(correct_prediction,tf.float32))
saver = tf.train.Saver()
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
for epoch in range(13):
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})
#print ("batch ",batch)
batch_xstest,batch_ystest = mnist.test.next_batch(batch_size)
# acc = sess.run(accuracy,feed_dict={x:mnist.test.images,y:mnist.test.labels,keep_prob:0.5})
acc = sess.run(accuracy,feed_dict={x: batch_xstest,y:batch_ystest,keep_prob:0.5})
print ("Iter "+ str(epoch) + ", Testing Accuracy= " + str(acc))
saver.save(sess,save_path='/mnist_net.ckpt')
结果
Iter 0, Testing Accuracy= 0.8517 Iter 1, Testing Accuracy= 0.9612 Iter 2, Testing Accuracy= 0.9769 Iter 3, Testing Accuracy= 0.9804 Iter 4, Testing Accuracy= 0.9832 Iter 5, Testing Accuracy= 0.9844 Iter 6, Testing Accuracy= 0.988 Iter 7, Testing Accuracy= 0.9882 Iter 8, Testing Accuracy= 0.9875 Iter 9, Testing Accuracy= 0.9889 Iter 10, Testing Accuracy= 0.9891 Iter 11, Testing Accuracy= 0.9897 Iter 12, Testing Accuracy= 0.9891 Iter 13, Testing Accuracy= 0.9897 Iter 14, Testing Accuracy= 0.9905 Iter 15, Testing Accuracy= 0.9913 Iter 16, Testing Accuracy= 0.9908 Iter 17, Testing Accuracy= 0.9909 Iter 18, Testing Accuracy= 0.9913 Iter 19, Testing Accuracy= 0.9915 Iter 20, Testing Accuracy= 0.9902 Iter 21, Testing Accuracy= 0.9899 Iter 22, Testing Accuracy= 0.9912 Iter 23, Testing Accuracy= 0.9911 Iter 24, Testing Accuracy= 0.9907 Iter 25, Testing Accuracy= 0.9918 Iter 26, Testing Accuracy= 0.9919 Iter 27, Testing Accuracy= 0.9916 Iter 28, Testing Accuracy= 0.9899 Iter 29, Testing Accuracy= 0.9924 Iter 30, Testing Accuracy= 0.9913 Iter 31, Testing Accuracy= 0.992 Iter 32, Testing Accuracy= 0.9927 Iter 33, Testing Accuracy= 0.9919 Iter 34, Testing Accuracy= 0.9922 Iter 35, Testing Accuracy= 0.9918 Iter 36, Testing Accuracy= 0.9932 Iter 37, Testing Accuracy= 0.9924 Iter 38, Testing Accuracy= 0.9917 Iter 39, Testing Accuracy= 0.9919 Iter 40, Testing Accuracy= 0.9933 Iter 41, Testing Accuracy= 0.9924 Iter 42, Testing Accuracy= 0.9926 Iter 43, Testing Accuracy= 0.9932 Iter 44, Testing Accuracy= 0.9922 Iter 45, Testing Accuracy= 0.9925 Iter 46, Testing Accuracy= 0.9928 Iter 47, Testing Accuracy= 0.9935 Iter 48, Testing Accuracy= 0.9922 Iter 49, Testing Accuracy= 0.9926