**
一 tf.concat( ) 函數–合并
**
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
|
In [2]: a = tf.ones([4,35,8]) In [3]: b = tf.ones([2,35,8]) In [4]: c = tf.concat([a,b],axis=0) In [5]: c.shape Out[5]: TensorShape([6, 35, 8])In [6]: a = tf.ones([4,32,8]) In [7]: b = tf.ones([4,3,8]) In [8]: c = tf.concat([a,b],axis=1) In [9]: c.shape Out[9]: TensorShape([4, 35, 8]) |
**
二 tf.stack( ) 函數–數據的堆疊,創建新的維度
**
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
|
In [2]: a = tf.ones([4,35,8]) In [3]: a.shape Out[3]: TensorShape([4, 35, 8])In [4]: b = tf.ones([4,35,8]) In [5]: b.shape Out[5]: TensorShape([4, 35, 8])In [6]: tf.concat([a,b],axis=-1).shape Out[6]: TensorShape([4, 35, 16])In [7]: tf.stack([a,b],axis=0).shape Out[7]: TensorShape([2, 4, 35, 8])In [8]: tf.stack([a,b],axis=3).shape Out[8]: TensorShape([4, 35, 8, 2]) |
**
三 tf.unstack( )函數–解堆疊
**
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
|
In [16]: a = tf.ones([4,35,8]) In [17]: b = tf.ones([4,35,8]) In [18]: c = tf.stack([a,b],axis=0) In [19]: a.shape,b.shape,c.shape Out[19]: (TensorShape([4, 35, 8]), TensorShape([4, 35, 8]), TensorShape([2, 4, 35, 8]))In [20]: aa,bb = tf.unstack(c,axis=0) In [21]: aa.shape,bb.shape Out[21]: (TensorShape([4, 35, 8]), TensorShape([4, 35, 8]))In [22]: res = tf.unstack(c,axis=1) In [23]: len(res) Out[23]: 4 |
**
四 tf.split( ) 函數
**
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
|
In [16]: a = tf.ones([4,35,8]) In [17]: b = tf.ones([4,35,8]) In [18]: c = tf.stack([a,b],axis=0) In [19]: a.shape,b.shape,c.shape Out[19]: (TensorShape([4, 35, 8]), TensorShape([4, 35, 8]), TensorShape([2, 4, 35, 8]))In [20]: aa,bb = tf.unstack(c,axis=0) In [21]: aa.shape,bb.shape Out[21]: (TensorShape([4, 35, 8]), TensorShape([4, 35, 8]))In [22]: res = tf.unstack(c,axis=1) In [23]: len(res) Out[23]: 4 |
以上這篇TensorFlow2.0:張量的合并與分割實例就是小編分享給大家的全部內容了，希望能給大家一個參考，也希望大家多多支持服務器之家。
原文鏈接：https://blog.csdn.net/meijie2018_1/article/details/99439186
