Tensorflow 保存和加载模型

import tensorflow as tf

def save_model():
    v1 = tf.Variable(tf.constant(1.0,shape=[1]),name='v1')
    v2 = tf.Variable(tf.constant(2.0,shape=[1]),name='v2')
    res = tf.add(v1,v2,name='add_res')
    saver = tf.train.Saver()
       
    with tf.Session() as sess:
        sess.run(tf.global_variables_initializer())
        saver.save(sess,'./save/model.ckpt')

def restore_model():
    saver = tf.train.import_meta_graph('./save/model.ckpt.meta')        
    with tf.Session() as sess:
        saver.restore(sess,'./save/model.ckpt')
        print (sess.run(tf.get_default_graph().get_tensor_by_name('add_res:0')))


if __name__=='__main__':
    
    save_mode() #第二次运行本程序时，需要注释这一句。
    restore_model()