import tensorflow as tf
import numpy as np

## Save to file
# remember to define the same dtype and shape when restore
W = tf.Variable([[1,2,3],[3,4,5]], dtype=tf.float32, name='weights')
b = tf.Variable([[1,2,3]], dtype=tf.float32, name='biases')

# init= tf.initialize_all_variables() # tf 马上就要废弃这种写法
# 替换成下面的写法:
init = tf.global_variables_initializer()
saver = tf.train.Saver()

with tf.Session() as sess:
    sess.run(init)
    save_path = saver.save(sess, "my_net/save_net.ckpt")
    print("Save to path: ", save_path)

Save to path:  my_net/save_net.ckpt

# 先建立 W, b 的容器
W = tf.Variable(np.arange(6).reshape((2, 3)), dtype=tf.float32, name="weights")
b = tf.Variable(np.arange(3).reshape((1, 3)), dtype=tf.float32, name="biases")

# 这里不需要初始化步骤 init= tf.initialize_all_variables()

saver = tf.train.Saver()
with tf.Session() as sess:
    # 提取变量
    saver.restore(sess, "my_net/save_net.ckpt")
    print("weights:", sess.run(W))
    print("biases:", sess.run(b))