This is a repo for my individual interest to understand Beta_vae, based on its ability to unentangle embedded representation. For orginal paper: beta-VAE: Learning Basic Visual Concepts with a Constrained Variational Framework https://openreview.net/forum?id=Sy2fzU9gl
More understanding of the theory: Understanding disentangling in β-VAE https://arxiv.org/abs/1804.03599
This version is the DeepMind of β-VAE.
The data utilized for β-VAE experimentation is from dsprites (from https://github.com/deepmind/dsprites-dataset).
Please init through python main.py
This display the location representation: Here is the orginal photo:
Here is the output of Beta_VAE photo: