python–tensorlow.expand_dims()
原文:https://www.geesforgeks.org/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附加维度。
语法: tensorflow.expand_dims(输入，轴，名称)
参数:
- 输入:是输入 Tensor。
- 轴:定义插入维度的索引。如果输入有 D 个尺寸，那么轴的值必须在[-(D+1)，D]范围内。
- 名称(可选):定义操作的名称。
返回:返回一个展开维度的张量。
例 1:
蟒蛇 3
# 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)
例 2:
蟒蛇 3
# 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)
