tensorflow基础知识9，tensorboard网络运行中各种数据显示

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

#载入数据
mnist=input_data.read_data_sets('mnist_data',one_hot=True)#noe_hot把像素点都转变成0或1的形式


#每个批次的大小，训练模型时，一次放入一批次
batch_size=100 #一批次100张图
#计算一共有多少个批次
n_batch=mnist.train.num_examples//batch_size# //是整除,得到批次数

#参数概要
def variable_summaries(var):#定义一个函数，作用是计算各种参数值
    with tf.name_scope('summaries'):
        mean=tf.reduce_mean(var)#计算平均值
        tf.summary.scalar('mean',mean)#记录平均值，将其命名为mean。summary.scalar用来显示标量信息
        with tf.name_scope('stddev'):
            stddev=tf.sqrt(tf.reduce_mean(tf.square(var-mean)))
        tf.summary.scalar('stddev',stddev)# 标准差
        tf.summary.scalar('max',tf.reduce_max(var))#最大值
        tf.summary.scalar('min',tf.reduce_min(var))#最小值
        tf.summary.histogram('histogram',var)#直方图

#命名空间
with tf.name_scope('input'):  #命名随意，比如input,下面的x和y要缩进，表示x，y放在input空间
#定义两个placeholder，配合上面命名空间，给x，y取个名字
   x=tf.placeholder(tf.float32,[None,784],name='x-input')#建立一个占位符，None是图片数，784是每幅图的像素个数
   y=tf.placeholder(tf.float32,[None,10],name='y-input')# 标签，建立一个占位符，10是指0-9十个数

with tf.name_scope('layer'):
#创建一个简单的神经网络，输入层784个神经元，输出层10个神经元，不设隐藏层
   with tf.name_scope('wights'):
      W=tf.Variable(tf.zeros([784,10]),name='W')#权值，设一个变量，置0
      variable_summaries(W)#把权值W当作参数，计算的各种指标
   with tf.name_scope('biases'):
      b=tf.Variable(tf.zeros([10]),name='b')#偏置值
      variable_summaries(b)#把偏置值b当作参数，计算的各种指标
   with tf.name_scope('wx_plus_b'):
      wx_plus_b=tf.matmul(x,W)+b
   with tf.name_scope('softmax'):
      prediction=tf.nn.softmax(tf.matmul(x,W)+b)#信号总和，经过softmax函数（**函数）转化成概率值

#二次代价函数
#loss =tf.reduce_mean(tf.square(y-prediction))

#使用交叉熵代价函数
with tf.name_scope('loss'):
    loss=tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=y,logits=prediction))
    tf.summary.scalar('loss',loss)

#使用梯度下降法
with tf.name_scope('train'):
    train_step=tf.train.GradientDescentOptimizer(0.2).minimize(loss)

#初始化变量
init=tf.global_variables_initializer()

with tf.name_scope('accuracy'):
    with tf.name_scope('correct_prediction'):
#训练好后求准确率，结果存放在一个布尔型列表中，argmax返回一维张量中最大的值所在的位置
        correct_prediction=tf.equal(tf.argmax(y,1),tf.argmax(prediction,1))#argmax函数是对行或列计算最大值，1表示按行，0表示按列，找到最大概率标签的位置。 equal函数是比较两个参数大小，相等的话返回True，不相等返回False
    with tf.name_scope('accuracy'):
#求准确率
        accuracy=tf.reduce_mean(tf.cast(correct_prediction,tf.float32))#cast()是类型转换函数，把布尔型参数转换为32位古典型,然后求平均值。true变成1.0，flse变成0
#Boolean→数值型：True转换为-1，False转换为0。数值型→Boolean：0转换为False，其他转换为True
        tf.summary.scalar('accuracy',accuracy)

#合并所有的summary,并将其加入到sess.run的语句里
merged=tf.summary.merge_all()

with tf.Session() as sess:
    sess.run(init)#初始化变量
    writer=tf.summary.FileWriter('logs/',sess.graph)#'logs/'是路径，graph存在logs文件夹中，如果没有logs文件夹，这里会自动生成
    for epoch in range(51):#迭代21个周期，把所有图片训练21次
        for batch in range(n_batch):
            batch_xs,batch_ys=mnist.train.next_batch(batch_size)#一次分配100张图片，图片数据保存在batch_xs，标签保存在batch_ys
            summary,_=sess.run([merged,train_step],feed_dict={x:batch_xs,y:batch_ys})#每tain训练一次，统计一次参数merged，运行后得到的merged存在summary里

        writer.add_summary(summary,epoch)#将summary和运行周期epoch写入tensorboard文件
        acc=sess.run(accuracy,feed_dict={x:mnist.test.images,y:mnist.test.labels})#x输入测试图片，从而得到prediciton的y，从而和label y 对比
        print('Iter'+str(epoch)+',Testing Accuracy'+str(acc))





Iter0,Testing Accuracy0.8248
Iter1,Testing Accuracy0.8922
Iter2,Testing Accuracy0.9016
Iter3,Testing Accuracy0.9059
Iter4,Testing Accuracy0.9083
Iter5,Testing Accuracy0.91
Iter6,Testing Accuracy0.9127
Iter7,Testing Accuracy0.9134
Iter8,Testing Accuracy0.9157
Iter9,Testing Accuracy0.9153
Iter10,Testing Accuracy0.9183
Iter11,Testing Accuracy0.9182
Iter12,Testing Accuracy0.9186
Iter13,Testing Accuracy0.9194
Iter14,Testing Accuracy0.9207
Iter15,Testing Accuracy0.9197
Iter16,Testing Accuracy0.9203
Iter17,Testing Accuracy0.9211
Iter18,Testing Accuracy0.9213
Iter19,Testing Accuracy0.922
Iter20,Testing Accuracy0.9223
Iter21,Testing Accuracy0.9214
Iter22,Testing Accuracy0.9219
Iter23,Testing Accuracy0.922
Iter24,Testing Accuracy0.9232
Iter25,Testing Accuracy0.9233
Iter26,Testing Accuracy0.9231
Iter27,Testing Accuracy0.9235
Iter28,Testing Accuracy0.9235
Iter29,Testing Accuracy0.9237
Iter30,Testing Accuracy0.9245
Iter31,Testing Accuracy0.9243
Iter32,Testing Accuracy0.9248
Iter33,Testing Accuracy0.9257
Iter34,Testing Accuracy0.9242
Iter35,Testing Accuracy0.9251
Iter36,Testing Accuracy0.9249
Iter37,Testing Accuracy0.9243
Iter38,Testing Accuracy0.9247
Iter39,Testing Accuracy0.9252
Iter40,Testing Accuracy0.9257
Iter41,Testing Accuracy0.9256
Iter42,Testing Accuracy0.9259
Iter43,Testing Accuracy0.9258
Iter44,Testing Accuracy0.9272
Iter45,Testing Accuracy0.9262
Iter46,Testing Accuracy0.9258
Iter47,Testing Accuracy0.9267
Iter48,Testing Accuracy0.9272
Iter49,Testing Accuracy0.927
Iter50,Testing Accuracy0.9273

 

 

运用上一节的办法打开tensorboard界面。