Sampling Disentangled Representation

[aissak21]

Is your feature request related to a problem? Please describe. Currently unable to sample across each node of the latent space to see what has been learned separately.

Describe the solution you'd like Implementation to see the independent features picked up by each node via sampling from the latent space. For example, trained Disentangled Beta VAE with a latent dimension of 16 and would like to see the features from each node independently.

Describe alternatives you've considered Sampling in random iterations and variety of samplers but unable to extend the capabilities of the modules any further.