numpy 除法时的broadcasting

import numpy as np

bbox_xywh = np.array([5,4,3,2])
strides = np.array([8, 16, 32])
bbox_xywh_scaled = 1.0 * bbox_xywh[np.newaxis, :] / strides[:, np.newaxis]
print('***************************')
print(bbox_xywh[np.newaxis, :])
print(strides[:,np.newaxis])
print('bbox_xywh[np.newaxis, :].shape=',bbox_xywh[np.newaxis, :].shape)
print('strides[:,np.newaxis].shape=',strides[:,np.newaxis].shape)
print(bbox_xywh_scaled.shape)
print(bbox_xywh_scaled)