跟我学算法- tensorflow模型的保存与读取 tf.train.Saver()

import tensorflow as tf

#创建变量
v1 = tf.Variable(tf.random_normal([1, 2], name=‘v1‘))
v2 = tf.Variable(tf.random_normal([2, 3], name=‘v2‘))
#初始化变量
init_op = tf.global_variables_initializer()
#构建训练模型的保存
saver = tf.train.Saver()
with tf.Session() as sess:
    sess.run(init_op)
    print(‘V1:‘, sess.run(v1))
    print(‘V2:‘, sess.run(v2))
    # saver.save(保存内容， 保存路径)
    saver_path = saver.save(sess, ‘save/model.ckpt‘)
    print(‘Model saved in file:‘, saver_path)

import tensorflow as tf
# v1，v2的设定，主要是看看输出的v1是哪个v1
v1 = tf.Variable(tf.random_normal([1, 2]), name=‘v1‘)
v2 = tf.Variable(tf.random_normal([2, 3]), name=‘v2‘)
# 构建保存模型
saver = tf.train.Saver()
with tf.Session() as sess:
    # 重新加载模型(重新赋予名字， 加载的路径)
    saver.restore(sess, ‘save/model.ckpt‘)
    print(‘V1:‘, sess.run(v1))
    print(‘V2:‘, sess.run(v2))
    print(‘Model restored‘)