Tensorflow 學習筆記 -----tf.where

TensorFlow函數：tf.where

在以前版本對應函數tf.select

官方解釋:

 1 tf.where(input, name=None)`
 2 Returns locations of true values in a boolean tensor.
 3 
 4 This operation returns the coordinates of true elements in input. The coordinates are returned in a 2-D tensor where the first dimension (rows) represents the number of true elements, and the second dimension (columns) represents the coordinates of the true elements. Keep in mind, the shape of the output tensor can vary depending on how many true values there are in input. Indices are output in row-major order.
 5 
 6 For example:
 7 # 'input' tensor is [[True, False]
 8 #                    [True, False]]
 9 # 'input' has two true values, so output has two coordinates.
10 # 'input' has rank of 2, so coordinates have two indices.
11 where(input) ==> [[0, 0],
12                   [1, 0]]
13 
14 # `input` tensor is [[[True, False]
15 #                     [True, False]]
16 #                    [[False, True]
17 #                     [False, True]]
18 #                    [[False, False]
19 #                     [False, True]]]
20 # 'input' has 5 true values, so output has 5 coordinates.
21 # 'input' has rank of 3, so coordinates have three indices.
22 where(input) ==> [[0, 0, 0],
23                   [0, 1, 0],
24                   [1, 0, 1],
25                   [1, 1, 1],
26                   [2, 1, 1]]

 

有兩種用法：

一、tf.where(tensor)

tensor 爲一個bool 型張量，where函數將返回其中爲true的元素的索引。如上圖官方註釋

二、tf.where(tensor,a,b)

a,b爲和ｔｅｎｓｏｒ相同維度的ｔｅｎｓｏｒ，將ｔｅｎｓｏｒ中的ｔｒｕｅ位置元素替換爲ａ中對應位置元素，ｆａｌｓｅ的替換爲ｂ中對應位置元素。

例：

import tensorflow as tf
import numpy as np
sess=tf.Session()

a=np.array([[1,0,0],[0,1,1]])
a1=np.array([[3,2,3],[4,5,6]])

print(sess.run(tf.equal(a,1)))
print(sess.run(tf.where(tf.equal(a,1),a1,1-a1)))

＞＞［［ｔｒｕｅ，ｆａｌｓｅ，ｆａｌｓｅ］，［ｆａｌｓｅ，ｔｒｕｅ，ｔｒｕｅ］］

＞＞［［３，-１，-２］，［-３，５，６］］