Beta-VAE Reproducibility: Challenges and Extensions

1 code implementation • 28 Dec 2021 • Miroslav Fil, Munib Mesinovic, Matthew Morris, Jonas Wildberger

$\beta$-VAE is a follow-up technique to variational autoencoders that proposes special weighting of the KL divergence term in the VAE loss to obtain disentangled representations.

Disentanglement