what is permutate argument 28perm 29 in tf transpose 3f

1x = tf.constant([[[ 1, 2, 3], 
2                    [ 4, 5, 6]], 
3                   [[ 7, 8, 9], 
4                    [ 10, 11, 12]], 
5                   [[ 13, 14, 15], 
6                    [ 16, 17, 18]], 
7                   [[ 19, 20, 21], 
8                    [ 22, 23, 24]]]) 
9transposed_tensor = tf.transpose(x, perm = [2,1,0]) 
10# if you run x.shape
11x.shape
12>>>TensorShape([4, 2, 3])
13# if you run transpose.shape
14transpose.shape
15>>>TensorShape([3,2,4])
16# now change the perm argument of tf.transpose to perm = [0,1,2]
17transpose.shape
18>>>TensorShape([4,2,3])
19## now change the perm argument of tf.transpose to perm = [1,2,0]
20transpose.shape
21>>>TensorShape([2,3,4])
22# that means if we consider the output of tensor as a list then we thought
23# list of perm argument as a index-list of shape of input array.so,if we change
24# the value of list in perm we see the corresponding change of input array shape
25# as output of transpose shape