(TensorFlow) - detailed explanation of tf.variable_scope and tf.name_scope

Learning tensorflow in the past few days, I have seen about tf.variable_scope and tf.name_scop, but I have not had a deep understanding of their functions.Reprint a blog:

The difference between tf.name_scope() and tf.variable_scope() in Tensorflow Record the effect,Forget one side later.

Briefly say it here:

tf.variable_scope allows variables to have the same name, including the variables obtained by tf.get_variable and the variables of tf.Variable

tf.name_scope allows variables to have the same name, only limited to tf.Variable variables

Example:

import tensorflow as tf;import numpy as np;import matplotlib.pyplot as plt;with tf.variable_scope('V1'):a1 = tf.get_variable(name='a1', shape=[1], initializer=tf.constant_initializer(1))a2 = tf.Variable(tf.random_normal(shape=[2,3], mean=0, stddev=1), name='a2')with tf.variable_scope('V2'):a3 = tf.get_variable(name='a1', shape=[1], initializer=tf.constant_initializer(1))a4 = tf.Variable(tf.random_normal(shape=[2,3], mean=0, stddev=1), name='a2')with tf.Session() as sess:sess.run(tf.initialize_all_variables())print a1.nameprint a2.nameprint a3.nameprint a4.name

Output:

V1/a1:0
V1/a2:0
V2/a1:0
V2/a2:0

Example 2:

import tensorflow as tf;import numpy as np;import matplotlib.pyplot as plt;with tf.name_scope('V1'):a1 = tf.get_variable(name='a1', shape=[1], initializer=tf.constant_initializer(1))a2 = tf.Variable(tf.random_normal(shape=[2,3], mean=0, stddev=1), name='a2')with tf.name_scope('V2'):a3 = tf.get_variable(name='a1', shape=[1], initializer=tf.constant_initializer(1))a4 = tf.Variable(tf.random_normal(shape=[2,3], mean=0, stddev=1), name='a2')with tf.Session() as sess:sess.run(tf.initialize_all_variables())print a1.nameprint a2.nameprint a3.nameprint a4.name

Error: Variable a1 already exists, disallowed. Did you mean to set reuse=True in VarScope? Originally defined at:

Replace with the following code to execute

import tensorflow as tf;import numpy as np;import matplotlib.pyplot as plt;with tf.name_scope('V1'):# a1 = tf.get_variable(name='a1', shape=[1], initializer=tf.constant_initializer(1))a2 = tf.Variable(tf.random_normal(shape=[2,3], mean=0, stddev=1), name='a2')with tf.name_scope('V2'):# a3 = tf.get_variable(name='a1', shape=[1], initializer=tf.constant_initializer(1))a4 = tf.Variable(tf.random_normal(shape=[2,3], mean=0, stddev=1), name='a2')with tf.Session() as sess:sess.run(tf.initialize_all_variables())# print a1.nameprint a2.name# print a3.nameprint a4.name

Output:

V1/a2:0
V2/a2:0