Can I use the VAE code in the example if I want a higher dimenional latent vector?

Question

The example is here: https://github.com/keras-team/keras/blob/master/examples/variational_autoencoder.py

As the question suggests, I just wanted to know whether just