执行回归后如何取消展平和成像？

Question

以下是我如何拼合这些图像：

N = images.shape[0]
images = np.reshape(images, (N, -1)) 
images_two = np.reshape(images_two, (N, -1))

我该如何撤销这个过程？

Answer 1

存储展平前的原始形状：

old_shape = images.shape
N = images.shape[0]
images = np.reshape(images, (N, -1)) 
images_two = np.reshape(images_two, (N, -1))

## to reshape

images.reshape(old_shape)

Answer 2

你可以这样做：

num_imgs = images.shape[0]
img_shape = images.shape[1:]

images_flat = images.reshape((num_imgs, -1))

images_unflat = images_flat.reshape((num_imgs, ) + img_shape))