How to sample from latent distribution

[crlz182]

Hello,

I was wondering how I can generate samples using the decoder network after training. In a VAE, I would just sample from the prior distribution z~N(0,1) and generate a data point using the decoder. In TriangularSylvesterVAE, however, I also have to provide hyperparameters lambda(x) that depend on the input. How can I sample from my latent distribution and generate samples from it?

I am new to normalizing flows in general and would appreciate any help.

[bfs18]

I am also curious about this.

[bgroenks96]

Just as in a regular VAE, the latent distribution is conditional on the input. Given an input x, you would obtain a mu and sigma from the encoder, and then draw samples from this latent distribution and feed them through the decoder.

It sounds like what you want is an unconditional prior p(z), which is not possible for a flow-based VAE unless you also define a prior on the flow parameters.