The generated images of the VAE depend on the latent space dimensions. The blurring reduces the dimension of the latent space is increased. The reconstructed images for 5-d, 10-d, and 20-d latent dimensions respectively, are as follows:
![](http://images-20200215.ebookreading.net/16/3/3/9781788293594/9781788293594__tensorflow-1x-deep__9781788293594__assets__7619532a-5c71-4c3f-9a89-4093fede393a.png)
The generated images of the VAE depend on the latent space dimensions. The blurring reduces the dimension of the latent space is increased. The reconstructed images for 5-d, 10-d, and 20-d latent dimensions respectively, are as follows:
3.144.172.223