tf.initialize_all_variables() 和 tf.initialize_local_variables() 有什么区别? [英] What is the difference between tf.initialize_all_variables() and tf.initialize_local_variables()?

问题描述

我正在查看此示例中的代码:fully_connected_reader.py

I am reviewing the code in this example: fully_connected_reader.py

我对第 147 和 148 行感到困惑:

I am confused with Line 147 and 148:

init_op = tf.group(tf.initialize_all_variables(),
                   tf.initialize_local_variables())

我不知道哪些变量是所有变量，哪些是局部变量.有什么想法吗?

I don't know which variables are all variables and which are local variables. Any ideas?

推荐答案

tf.initialize_all_variables() 是 tf.initialize_variables(tf.all_variables()) 的快捷方式, tf.initialize_local_variables() 是 tf.initialize_variables(tf.local_variables()) 的快捷方式，它在 GraphKeys.VARIABLES 和GraphKeys.LOCAL_VARIABLE 集合，分别.

tf.initialize_all_variables() is a shortcut to tf.initialize_variables(tf.all_variables()), tf.initialize_local_variables() is a shortcut to tf.initialize_variables(tf.local_variables()), which initializes variables in GraphKeys.VARIABLES and GraphKeys.LOCAL_VARIABLE collections, respectively.

GraphKeys.LOCAL_VARIABLES 集合中的变量是添加到图中但未保存或恢复的变量 (source).

Variables in GraphKeys.LOCAL_VARIABLES collection are variables that are added to the graph, but not saved or restored (source).

tf.Variable() 默认为 GraphKeys.VARIABLE 集合添加一个新变量，可以通过 collections= 参数控制.

tf.Variable() by default adds a new variable to GraphKeys.VARIABLE collection, which can be controlled by collections= argument.

这篇关于tf.initialize_all_variables() 和 tf.initialize_local_variables() 有什么区别?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！