TensorFlow保存TensorBoard图像操作

简单的代码：

import tensorflow as tf

In [2]: matrix1=tf.constant([[3.,3.]])
In [3]: matrix2=tf.constant([[2.],[2.]])

with tf.Session() as sess:
  ...:   writer = tf.summary.FileWriter('./graph', sess.graph)
  ...:   result = sess.run(tf.matmul(matrix1, matrix2))
  ...:   writer.close()

ipython中使用！+命令可以直接运行terminal命令。

terminal输入： tensorboard --logdir graph/

跳出：Starting TensorBoard 54 at http://amax:6006

在浏览器输入地址加端口号并在graph中查看。

补充知识：tensorflow 利用保存的meta图文件生成log供tensorboard可视化保存恢复模型

tensorboard可视化图：

import tensorflow as tf
g = tf.Graph() 
with g.as_default() as g: 
  tf.train.import_meta_graph('criteo_80.meta') 
 
with tf.Session(graph=g) as sess: 
  file_writer = tf.summary.FileWriter(logdir='./', graph=g)

保存恢复模型：

# 建模型
saver = tf.train.Saver()
 
with tf.Session() as sess:
  # 存模型，注意此处的model是文件名非路径
  saver.save(sess, "/tmp/model")
 
with tf.Session() as sess:
  # 恢复模型
  saver.restore(sess, "/tmp/model")

# 先恢复图
saver = tf.train.import_meta_graph("/tmp/model.meta")
 
with tf.Session() as sess:
  # 再恢复参数
  saver.restore(sess, "/tmp/model")

频道导航

TensorFlow保存TensorBoard图像操作

您可能感兴趣的文章: