This repository contains the code for
-
the empirical study about LIDVAE on images and text (Section 4.1)
-
the empirical study about latent variable non-identifiability and posterior collapse in PPCA (Section 4.2)
The code is tested in Python 3.7.3, Pytorch 1.4.0, and Tensorflow 1.14.0.
Wang, Y., Blei, D.M., and Cunningham, J.P. (2021) Posterior Collapse and Latent Variable Non-identifiability. NeurIPS 2021. [link]