tf.stack() tf.unstack() tf.concat() tf.split()

目录

拼接操作: (tf.concat,tf.stack)

分割操作: (tf.unstack,tf.split) 

对比总结:

---

 

a = tf.constant([[1,2,3],[3,4,5]]) # shape (2,3)
b = tf.constant([[7,8,9],[10,11,12]]) # shape (2,3)

拼接操作: (tf.concat,tf.stack)

ab1 = tf.concat([a,b], axis=0) # shape(4,3)
[[ 1  2  3]
 [ 3  4  5]
 [ 7  8  9]
 [10 11 12]]

1. tf.stack其作用类似于tf.concat，都是拼接两个张量，
2. 不同之处，tf.concat拼接的是除了拼接维度axis外, 其他维度的shape完全相同的张量，并且产生的张量的阶数不会发生变化，而tf.stack则会在新的张量阶上拼接，产生的张量的阶数将会增加。

 

分割操作: (tf.unstack,tf.split) 

tf.unstack

import tensorflow as tf
a = tf.constant([[1, 2, 3], [4, 5, 6]])
b = tf.unstack(a, axis=0)
c = tf.unstack(a, axis=1)
with tf.Session() as sess:
        print(sess.run(a))
        print(sess.run(b))
        print(sess.run(c))

输出

[[1 2 3]
 [4 5 6]]

[array([1, 2, 3], dtype=int32), array([4, 5, 6], dtype=int32)]

[array([1, 4], dtype=int32), array([2, 5], dtype=int32), array([3, 6], dtype=int32)]

 

tf.split

import tensorflow as tf

a = tf.constant([[1, 2, 3], [4, 5, 6]])
b = tf.split(a,num_or_size_splits=2,axis=0)

with tf.Session() as sess:
    print(sess.run(a))
    print(sess.run(b))

[[1 2 3]
 [4 5 6]]
[array([[1, 2, 3]]), array([[4, 5, 6]])]

 

对比总结:

import tensorflow as tf

A = [[1, 2, 3], [4, 5, 6]]
a0 = tf.split(A, num_or_size_splits=3, axis=1)#不改变维数（！！）
a1 = tf.unstack(A, num=3,axis=1)
a2 = tf.split(A, num_or_size_splits=2, axis=0)
a3 = tf.unstack(A, num=2,axis=0)
with tf.Session() as sess:
    # print(sess.run(a0))
    # print(sess.run(a1))
    print(sess.run(a2))
    # print(sess.run(a3)