【Tensorflow】Tensorboard最简版本使用教程

参考文章:https://www.cnblogs.com/fydeblog/p/7429344.html

本代码在【Tensorflow】调用已经保存好的模型继续训练(最简版)基础上添加Tensorboard必要的代码

1.显示网络结构图

import tensorflow as tf
from pathlib import Path

def save(path='./model/transfer_learn'):
    saver = tf.train.Saver()
    saver.save(sess, path, write_meta_graph=False)

def restore(path='./model/transfer_learn'):
    saver = tf.train.Saver()
    saver.restore(sess, path)

def train(input):
    a = tf.assign_add(input, tf.ones(1))
    return a

if __name__ == "__main__":
	a = tf.Variable(tf.zeros(1),name='a')
    model_file = Path('./model')

    with tf.Session() as sess:
        writer = tf.summary.FileWriter("logs/", sess.graph)

        if model_file.exists():
            restore()
        else:
            sess.run(tf.initialize_all_variables())
        for i in range(5):
            output = train(a)
            print(sess.run(output))
        save()

在Session内添加writer = tf.summary.FileWriter("logs/", sess.graph),会生成logs文件夹.
打开终端,不需要进到logs内,输入tensorboard --logdir logs,然后打开生成的链接即可.