卷积神经网络的训练和测试（针对电脑内存比较小的，运行速度比较慢的）

#!/usr/bin/env python

# -*- coding: utf-8 -*-

from tensorflow.examples.tutorials.mnist import input_data

import tensorflow as tf

import numpy as np

mnist = input_data.read_data_sets("MNIST_data/", one_hot=True)

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')

def my_conv(input_image, out_dim, name,channel):

    with tf.variable_scope(name):

        w_conv1 = weight_variable([5, 5, channel, out_dim])

        b_conv1 = bias_variable([out_dim])

        h_conv1 = tf.nn.relu(conv2d(input_image, w_conv1) + b_conv1)

        h_pool1 = max_pool_2x2(h_conv1)

    return h_pool1

def my_fc_layer(input_image, out_dim, name):

    with tf.variable_scope(name):

        w_fc1 = weight_variable([7*7* 64, out_dim])

        b_fc1 = bias_variable([out_dim])

        h_pool2_flat = tf.reshape(input_image, [-1,7 *7 * 64])

        h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, w_fc1) + b_fc1)

    return h_fc1

def build_net(input_data, keep_prob):

    conv1 = my_conv(input_image=input_data, out_dim=32, name='conv_layer1',channel=1)

    conv2 = my_conv(input_image=conv1, out_dim=64, name='conv_layer2',channel=32)

    fc1 = my_fc_layer(input_image=conv2, out_dim=1024, name='fc_layer1')

    h_fc1_drop = tf.nn.dropout(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)

    return  y_conv

x = tf.placeholder(tf.float32, [None, 784])

y_ = tf.placeholder(tf.float32, [None, 10])

x_image = tf.reshape(x, [-1, 28, 28, 1])

sum=tf.Variable(0.0,name="sum")

temp=tf.Variable(0.0,name="temp")

keep_prob = tf.placeholder(tf.float32)

y_conv=build_net(input_data=x_image,keep_prob=keep_prob)

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)

correct_prediction = tf.equal(tf.argmax(y_conv, 1), tf.argmax(y_, 1))

accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))

tf.global_variables_initializer().run()  # 启动Session

for i in range(500):

    batch = mnist.train.next_batch(50)

    if i % 100 == 0:

        train_accuracy = accuracy.eval(feed_dict={x: batch[0], y_: batch[1], keep_prob: 1.0})

        print("step %d, training accuracy %g" % (i, train_accuracy))

    train_step.run(feed_dict={x: batch[0], y_: batch[1], keep_prob: 0.5})

#下面是训练时发现电脑内存较小，只能把训练集拆分成多步完成

for i in range(200):

    testSet = mnist.test.next_batch(50)

    temp=accuracy.eval(feed_dict={ x: testSet[0], y_: testSet[1], keep_prob: 1.0})

    #print("test accuracy %g"%accuracy.eval(feed_dict={ x: testSet[0], y_: testSet[1], keep_prob: 1.0}))

    print("test accuracy %g"%temp)

    sum= tf.add(sum , temp)

s=sess.run(sum)

print(s/200)