tf.reverse

在之前的博客中我说过，

一维度的矩阵是队列

二维度的矩阵是方阵

三维度的矩阵是大楼

四维度的矩阵是小区

我以三维矩阵来说 tf.reverse的意义，可以把它想象成一座大楼

import tensorflow  as tf 
t=tf.constant( [ [[ 0,  1,  2,  3],
                  [ 4,  5,  6,  7],
                  [ 8,  9, 10, 11]],
                [[ 10,  11,  12,  13],
                  [ 14,  15,  16,  17],
                  [ 18,  19, 20, 21]]
                
                ])

tf.reverse(t, [0]) 就是楼层互换，顶楼到底楼

tf.reverse(t, [0])
Out[55]: 
<tf.Tensor: shape=(2, 3, 4), dtype=int32, numpy=
array([[[10, 11, 12, 13],
        [14, 15, 16, 17],
        [18, 19, 20, 21]],

       [[ 0,  1,  2,  3],
        [ 4,  5,  6,  7],
        [ 8,  9, 10, 11]]], dtype=int32)>

tf.reverse(t, [1]) 每层的列倒序

tf.reverse(t, [1])
Out[56]: 
<tf.Tensor: shape=(2, 3, 4), dtype=int32, numpy=
array([[[ 8,  9, 10, 11],
        [ 4,  5,  6,  7],
        [ 0,  1,  2,  3]],

       [[18, 19, 20, 21],
        [14, 15, 16, 17],
        [10, 11, 12, 13]]], dtype=int32)>

tf.reverse(t, [1]) 每层的倒行序

tf.reverse(t, [2])
Out[57]: 
<tf.Tensor: shape=(2, 3, 4), dtype=int32, numpy=
array([[[ 3,  2,  1,  0],
        [ 7,  6,  5,  4],
        [11, 10,  9,  8]],

       [[13, 12, 11, 10],
        [17, 16, 15, 14],
        [21, 20, 19, 18]]], dtype=int32)>