Tensorflow项目实战一：MNIST手写数字识别

Posted 2020-11-03 HOU_JUN

　　此模型中，输入是28*28*1的图片，经过两个卷积层（卷积+池化）层之后，尺寸变为7*7*64，将最后一个卷积层展成一个以为向量，然后接两个全连接层，第一个全连接层加一个dropout，最后一个全连接层输出10个分类的预测结果，然后计算损失，进行训练。

　　代码如下：

import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data

#定义一个获取卷积核的函数
def weight_variable(shape):
    initial = tf.truncated_normal(shape, stddev=0.1)
    return tf.Variable(initial)

#定义一个获取偏置值的函数
def bias_variable(shape):
    initial = tf.constant(0.1,shape=shape)
    return tf.Variable(initial)

#定义一个卷积函数
def conv2d(x,W):
    return tf.nn.conv2d(x,W,[1,1,1,1],padding="SAME")

#定义一个池化函数
def max_pool_2x2(x):
    return tf.nn.max_pool(x,ksize=[1,2,2,1], strides=[1,2,2,1],padding="VALID")


if __name__ == "__main__":
    mnist = input_data.read_data_sets("MNIST_data/",one_hot=True)
    x = tf.placeholder(shape=[None,28*28],dtype=tf.float32)
    lable = tf.placeholder(shape=[None,10],dtype=tf.float32)

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

    #第一个卷积层
    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)
    #14*14*32

    #第二个卷积层
    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)
    #7*7*64

    #全连接层，输出为1024维向量
    W_fc1 = weight_variable([7*7*64,1024])
    b_fc1 = weight_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_dropout = tf.nn.dropout(h_fc1,keep_prob=keep_prob)

    #把1024维向量转换成10维，对应10个类别
    W_fc2 = weight_variable([1024,10])
    b_fc2 = weight_variable([10])
    y_conv = tf.matmul(h_fc1,W_fc2)+b_fc2

    #直接使用tf.nn.softmax_cross_entropy_with_logits直接计算交叉熵
    cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=lable,logits=y_conv))
    #定义train_step
    train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)

    #定义测试的准确率
    correct_prediction = tf.equal(tf.argmax(y_conv,1),tf.argmax(lable,1))
    accuracy = tf.reduce_mean(tf.cast(correct_prediction,tf.float32))

    # 创建Session和变量初始化
    sess = tf.InteractiveSession()
    sess.run(tf.global_variables_initializer())

    #训练20000步
    for i in range(20000):
        batch = mnist.train.next_batch(50)
        if i % 100==0:
            train_accuracy = sess.run(accuracy,feed_dict={
                x:batch[0],lable:batch[1],keep_prob: 1.0})
            print("step %d, training accuracy %g" % (i, train_accuracy))
        _ = sess.run(train_step, feed_dict={x: batch[0], lable: batch[1], keep_prob: 0.5})
    print("test accuracy %g" % sess.run(accuracy, feed_dict={
        x: mnist.test.images, lable: mnist.test.labels, keep_prob: 1.0}))