1、启动方法:
运行已经训练好的代码
使用命令
tensorboard --logdir=C:\Users\vip_g\logs
2、要重新编译代码、更新图怎么做?
1)要去logs目录下把文件删除
2)运行代码
3)重新执行启动tensorboard命令
4)刷新tensorboard界面
3、TensorBoard的使用流程
(1)添加记录节点:tf.summary.scalar/image/histogram()等
比如 : tf.summary.scalar(“accuarcy_train”, accuracy_train) 对精确度的统计,第一个参数是名字,第二个参数是变量名
(2)汇总记录节点:merged = tf.summary.merge_all()
(3)运行汇总节点:summary = sess.run(merged),得到汇总结果
(4)日志书写器实例化:summary_writer = tf.summary.FileWriter(logdir, graph=sess.graph),实例化的同时传入 graph 将当前计算图写入日志
(5)调用日志书写器实例对象summary_writer的add_summary(summary, global_step=i)方法可依次将所有汇总日志写入文件:
把第i次迭代的结果summary添加到train_writer: summary_writer.add_summary(summary, i)
(6)调用日志书写器实例对象summary_writer的close()方法写入内存,否则它每隔120s写入一次
summary_writer.close()
4、举例
import tensorflow as tf
#定义命名空间
# with tf.name_scope('input'):
# #fetch:就是同时运行多个op的意思
# input1 = tf.constant(3.0,name='A')#定义名称,会在tensorboard中代替显示
# input2 = tf.constant(4.0,name='B')
# input3 = tf.constant(5.0,name='C')
# with tf.name_scope('op'):
# #加法
# add = tf.add(input2,input3)
# #乘法
# mul = tf.multiply(input1,add)
# with tf.Session() as ss:
# #默认在当前py目录下的logs文件夹,没有会自己创建
# wirter = tf.summary.FileWriter('logs/',ss.graph)
# result = ss.run([mul,add])
# print(result)
with tf.name_scope('input'):
input1=tf.constant(3.0,name='A')
input2=tf.constant(4.0,name='B')
input3=tf.constant(5.0,name='C')
with tf.name_scope('op'):
#加法
add=tf.add(input2,input3)
#乘法
mul=tf.multiply(input1,add)
with tf.Session() as ss:
writer=tf.summary.FileWriter('logs/',ss.graph)
result=ss.run([mul,add])
print(result)
参考文章:
https://www.cnblogs.com/baby-lily/p/10931302.html
https://www.jianshu.com/p/2f1bc36c08e0
主要代码:
merged_summary_op = tf.summary.merge_all() # 合并所有的summary
summary_wirter = tf.summary.FileWriter(‘log/闪电near’, sess.graph)
tf.summary.scalar(‘loss’, cost)
summary_str = sess.run(merged_summary_op, feed_dict={X:train_x, Y:train_y})
summary_wirter.add_summary(summary_str, epoch)