深度卷积：pool时feature为奇数处理方法："VALID" to "SAME"

最新推荐文章于 2022-05-22 22:01:30 发布

取名字什么玩意儿

最新推荐文章于 2022-05-22 22:01:30 发布

阅读量929

点赞数

文章标签：卷积神经网络

本文链接：https://blog.csdn.net/weixin_44011829/article/details/105251514

版权

import tensorflow as tf
input_tensor=tf.get_variable(name="input",shape=[1,5,5,1],initializer=tf.truncated_normal_initializer(stddev=0.1))
out=tf.nn.max_pool(input_tensor,ksize=[1,2,2,1],strides=[1,2,2,1],padding="SAME")

with tf.Session() as sess:
    init = tf.global_variables_initializer()
    sess.run(init)
    print(sess.run(input_tensor))
    print("===========================")
    cur_out=sess.run(out)
    print(cur_out)

putout:
[[[[ 0.01732349]
   [ 0.0183203 ]
   [ 0.02233031]
   [-0.15542398]
   [-0.0119805 ]]

  [[ 0.02118089]
   [ 0.01722614]
   [-0.02548178]
   [-0.0288796 ]
   [ 0.19945478]]

  [[ 0.09523761]
   [-0.0492761 ]
   [-0.11030612]
   [-0.03635116]
   [ 0.02534419]]

  [[ 0.12810625]
   [-0.17728545]
   [-0.00613894]
   [ 0.13418455]
   [ 0.05811666]]

  [[ 0.07808375]
   [ 0.01789804]
   [ 0.14812444]
   [-0.07404469]
   [ 0.1617347 ]]]]
===========================
[[[[0.02118089]
   [0.02233031]
   [0.19945478]]

  [[0.12810625]
   [0.13418455]
   [0.05811666]]

  [[0.07808375]
   [0.14812444]
   [0.1617347 ]]]]

Process finished with exit code 0