Variational Auto-encoder with Non-parametric Bayesian Prior
Instruction:
use vae_*.py to train the model.
For test and analysis you can use DPVAE.r.
Examples:
source('DPVAE.r')
init('model')
z=sample(10)
image(display(z)[1,,])
image(display(z)[2,,])
image(display(z)[3,,])
image(display(z)[8,,])
z2=reconstruct(z)
image(display(z2)[8,,])
z2=reconstruct(z)
image(display(z2)[8,,])
image(display(z)[8,,])
image(display(z2)[8,,])
z2=reconstruct(z)
image(display(z2)[8,,])
z2=reconstruct(z)
image(display(z2)[8,,])
z2=reconstruct(z)
image(display(z2)[8,,])
z2=reconstruct(z)
image(display(z2)[8,,])
z2=reconstruct(z)
image(display(z2)[8,,])
z2=reconstruct(z)
image(display(z2)[8,,])