概述
tf.newaxis的功能与np.newaxis的功能、用法相同，是增加维度的
与tf.expand_dims()功能相同，用法不同
例子
对一维数组改变维度
a=tf.Variable([1,2,3,4])
print("a.shape:",a.shape)
print("a:",a)
print("-"*50)
a2 = a[tf.newaxis, :]
print("a2.shape:",a2.shape)
print("a2:",a2)
print("-"*50)
a3 = a[:,tf.newaxis]
print("a3.shape:",a3.shape)
print("a3:",a3)
执行结果：
a.shape: (4,)
a: <tf.Variable 'Variable:0' shape=(4,) dtype=int32, numpy=array([1, 2, 3, 4])>
--------------------------------------------------
a2.shape: (1, 4)
a2: tf.Tensor([[1 2 3 4]], shape=(1, 4), dtype=int32)
--------------------------------------------------
a3.shape: (4, 1)
a3: tf.Tensor(
[[1]
[2]
[3]
[4]], shape=(4, 1), dtype=int32)
例子2
对二维数组改变维度
在原有的2个维度中间插入一个维度
b=tf.constant([[1,2,3],[4,5,6]])
print("b.shape:",b.shape)
print("b:",b)
print("-"*50)
b2 = b[:,tf.newaxis]
print("b2.shape:",b2.shape)
print("b2:",b2)
在原有的维度后增加一个维度
b=tf.constant([[1,2,3],[4,5,6]])
print("b.shape:",b.shape)
print("b:",b)
print("-"*50)
b3 = b[:,:,tf.newaxis]
print("b3.shape:",b3.shape)
print("b3:",b3)