cnn.py

# -*- coding: utf-8 -*-
# mnist 数据集




import tensorflow as tf
import tensorflow.examples.tutorials.mnist.input_data as input_data
mnist = input_data.read_data_sets("MNIST_data/", one_hot=True)#下载数据集
x= tf.placeholder(tf.float32,[None, 784])#x是n*（28*28）
y_actual=tf.placeholder(tf.float32,shape=[None,10])#输出10维


def weight_variable(shape):#初始化weight
    initial=tf.truncated_normal(shape, stddev=0.1)
    return tf.Variable(initial)

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

def conv2d(x, W):
  return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')

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

#-----------------------------------------------------------------------------------------------#
x_image = tf.reshape(x, [-1,28,28,1])         #转换输入数据shape,以便于用于网络中
W_conv1 = weight_variable([5, 5, 1, 6])      
b_conv1 = bias_variable([6])       
h_conv1 = tf.nn.softmax(conv2d(x_image, W_conv1) + b_conv1)     #第一个卷积层
h_pool1 = max_pool(h_conv1)                                  #第一个池化层

W_conv2=weight_variable([5,5,6,12])
b_conv2=bias_variable([12])
h_conv2=tf.nn.softmax(conv2d(h_pool1,W_conv2)+b_conv2)
h_pool2=max_pool(h_conv2)

W_fc1=weight_variable([7*7*12,1024])
b_fc1=bias_variable([1024])
h_pool2_flat=tf.reshape(h_pool2,[-1,7*7*12])
h_fc1=tf.nn.relu(tf.matmul(h_pool2_flat,W_fc1)+b_fc1)

keep_prob=tf.placeholder("float")
h_fc1_drop=tf.nn.dropout(h_fc1,keep_prob)

W_fc2=weight_variable([1024,10])
b_fc2=bias_variable([10])
y_predict=tf.nn.softmax(tf.matmul(h_fc1_drop,W_fc2)+b_fc2)
#--------------------------------------正向传播------------------------------------------#


cross_entropy = -tf.reduce_sum(y_actual*tf.log(y_predict))#计算交叉熵，以后在弄懂，这里只讲实现
train_step = tf.train.GradientDescentOptimizer(1e-3).minimize(cross_entropy)    #梯度下降法
correct_prediction = tf.equal(tf.argmax(y_predict,1), tf.argmax(y_actual,1))    
accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float"))                 #精确度计算
sess=tf.InteractiveSession()                          
sess.run(tf.initialize_all_variables())
for i in range(200000):
  batch = mnist.train.next_batch(500)
  if i%100 == 0:                  #训练100次，验证一次
    test_acc=accuracy.eval(feed_dict={x: mnist.test.images, y_actual: mnist.test.labels, keep_prob: 1.0})
    print ("test accuracy",test_acc)
    #train_acc = accuracy.eval(feed_dict={x:batch[0], y_actual: batch[1], keep_prob: 1.0})
    #print('step',i,'training accuracy',train_acc)
    train_step.run(feed_dict={x: batch[0], y_actual: batch[1], keep_prob: 0.8})

test_acc=accuracy.eval(feed_dict={x: mnist.test.images, y_actual: mnist.test.labels, keep_prob: 1.0})
print ("test accuracy",test_acc)