一、概要

基於Tensorflow 1.0+版本實現，利用mnist資料集訓練CNN，達到了99.6%的準確率。

二、CNN結構

1.兩個卷積層、兩個池化層、一個全連線層、一個Dropout層以及一個Softmax層。
2.原始資料為28*28的大小、單通道的圖片。
3.第一個卷積層：5*5的卷積核，1個通道，32個不同的卷積核；第一個池化層：2*2的最大池化。
4.第二個卷積層：5*5的卷積核，32個通道，64個不同的卷積核；第二個池化層：2*2的最大池化。
5.全連線層：1024個隱含節點。
6.Droupout層：隨機丟棄一部分節點資料避免過擬合。
7.Softmax層：最後的概率輸出。

三、實現

# -*- 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)

# print mnist.train.images.shape,mnist.train.labels.shape
sess = tf.InteractiveSession()

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)

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'
 
)

train_x = tf.placeholder(tf.float32,[None,784])
train_y = tf.placeholder(tf.float32,[None,10])
x_image = tf.reshape(train_x,[-1,28,28,1])

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)

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)

W_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)

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

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)

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

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

tf.global_variables_initializer().run()
for i in xrange(20000):
    batch = mnist.train.next_batch(50)
    if i%100 == 0:
        train_accuracy = accuracy.eval(feed_dict={train_x:batch[0],train_y:batch[1],keep_prob:1.0})
        print 'step %d,training accuracy %g' % (i,train_accuracy)
    train_step.run(feed_dict={train_x:batch[0],train_y:batch[1],keep_prob:0.5})

print 'test accuracy %g' % accuracy.eval(feed_dict={train_x:mnist.test.images,train_y:mnist.test.labels,keep_prob:1.0})