Reconstruct a image by VAE

Here, I build a VAE model by Pytorch. In general, one image input can generate one decoded output image. Can one image reconstruct more than one image, such as 5?

If this is possible, how can I do it with code?