Broadcasting可以理解成把维度分成大维度和小维度，小维度较为具体，大维度更加抽象。也就是小维度针对某个示例，然后让这个示例通用语大维度。

import tensorflow as tf

x = tf.random.normal([4,32,32,3])

x.shape

(x+tf.random.normal([3])).shape

(x+tf.random.normal([32,32,1])).shape

(x+tf.random.normal([4,1,1,1])).shape

try:

    (x+tf.random.normal([1,4,1,1])).shape

except Exception as e:

    print(e)

(x+tf.random.normal([4,1,1,1])).shape

b = tf.broadcast_to(tf.random.normal([4,1,1,1]),[4,32,32,3])
b.shape

a = tf.ones([3,4])

a.shape

a1 = tf.broadcast_to(a,[2,3,4])

a1.shape

a2 = tf.expand_dims(a,axis=0)  # 0前插入一维

a2.shape

a2 = tf.tile(a2,[2,1,1])  # 复制一维2次，复制二、三维1次

a2.shape