tf实现LeNet5

加载数据

# encoding: utf-8
# author : wht@92
# LeNet5的tf的实现
# 2019-11-28
from tensorflow.examples.tutorials.mnist import input_data
import tensorflow as tf 
mnist = input_data.read_data_sets('MNIST_data/', one_hot=True)
sess = tf.InteractiveSession()

定义权重和偏置函数

def weight_variable(shape):
 initial = tf.truncated_normal(shape, stddev=0.1)
 return tf.Variable(initial)
 
def bias_variable(shape):
 initial = tf.constant(0.1, shape=shape)
 return tf.Variable(initial)

定义卷积层函数-卷积操作和池化操作

def conv2d(x,W):
 return tf.nn.conv2d(x, W, strides=[1,1,1,1], padding='SAME')
def max_pool_2x2(x):
 return tf.nn.max_pool(x, ksize=[1,2,2,1], strides=[1,2,2,1], padding='SAME')

定义数据格式
将输入的1D数据转换为2D数据，[-1,28,28,1] ，-1代表样本数量不定，输入为7841，转换成2828的数据，1代表单颜色通道

#定义占位符
x = tf.placeholder(tf.float32,[None, 784])
y_ = tf.placeholder(tf.float32, [None, 10])
x_image = tf.reshape(x, [-1,28,28,1])

数据初始化，及第一层卷积
卷积核为55 ，1颜色通道，数量为32
激活函数为relu函数
最大池化22

#初始化第一层卷积层
W_conv1 = weight_variable([5,5,1,32])
b_conv1 = bias_variable([32])
h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1)+b_conv1)
h_pool1 = max_pool_2x2(h_conv1)

第二层卷积
卷积核5*5，数量64

W_conv2 = weight_variable([5,5,32,64])
b_conv2 = bias_variable([64])
h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2)+b_conv2)
h_pool2 = max_pool_2x2(h_conv2)

全连接层

_fc1 = weight_variable([7*7*64, 1024])
b_fc1 = bias_variable([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)

为减轻过拟合加入dropout层

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

softmax层

W_fc2 = weight_variable([1024,10])
b_fc2 = bias_variable([10])
y_conv = tf.nn.softmax(tf.matmul(h_fc1_drop, W_fc2)+b_fc2)

定义损失函数

cross_entropy = tf.reduce_mean(-tf.reduce_sum(y_*tf.log(y_conv), reduction_indices=[1]))
train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)

定义评测函数

correct_prediction = tf.equal(tf.argmax(y_conv,1), tf.argmax(y_,1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction,tf.float32))

训练时Dropout设置为0.5，批次设置为50，迭代20000次，每100次进行一次评测（评测设置Dropout为1.0）

tf.global_variables_initializer().run()
for i in range(20000):
 batch = mnist.train.next_batch(50)
 if i%100==0:
  trian_accuracy = accuracy.eval(feed_dict={x:batch[0], y_:batch[1], keep_prob:1.0})
  print("step %d, training accuracy %g"%(i, trian_accuracy))
 train_step.run(feed_dict={x:batch[0], y_:batch[1], keep_prob: 0.5})

最终评测数据

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