Python – tensorflow.expand_dims()

TensorFlow是谷歌设计的开源Python库，用于开发机器学习模型和深度学习神经网络。

expand_dims()用于在输入张量中插入一个额外的维度。

语法: tensorflow.expand_dims( input, axis, name)

参数:

input: 它是输入的张量。
axis: 它定义了要插入的尺寸的索引。如果输入有D个维度，那么轴的值必须在[-(D+1), D]范围内。
name（可选）:它定义了该操作的名称。

返回：它返回一个扩展维度的张量。

示例 1:

# Importing the library
import tensorflow as tf
 
# Initializing the input
x = tf.constant([[2, 3, 6], [4, 8, 15]])
 
# Printing the input
print('x:', x)
 
# Calculating result
res = tf.expand_dims(x, 1)
 
# Printing the result
print('res: ', res)

输出:

x: tf.Tensor(
[[ 2  3  6]
 [ 4  8 15]], shape=(2, 3), dtype=int32)
res:  tf.Tensor(
[[[ 2  3  6]]

 [[ 4  8 15]]], shape=(2, 1, 3), dtype=int32)

# shape has changed from (2, 3) to (2, 1, 3)

示例 2:

# Importing the library
import tensorflow as tf
 
# Initializing the input
x = tf.constant([[2, 3, 6], [4, 8, 15]])
 
# Printing the input
print('x:', x)
 
# Calculating result
res = tf.expand_dims(x, 0)
 
# Printing the result
print('res: ', res)

输出:

x: tf.Tensor(
[[ 2  3  6]
 [ 4  8 15]], shape=(2, 3), dtype=int32)
res:  tf.Tensor(
[[[ 2  3  6]
  [ 4  8 15]]], shape=(1, 2, 3), dtype=int32)

# shape has changed from (2, 3) to (1, 2, 3)