A tensorflow 2.0 implementation of the FactorVAE algorithm. (Not original implementation)
Disentangling by Factorising (Kim & Mnih, 2018) https://arxiv.org/pdf/1802.05983.pdf
You will need to download the file 3dshapes.h5 from https://github.com/deepmind/3d-shapes.
To train the model, run FactorVAE.ipynb
Results of latent traversal. Each column indicates individual latent dimensions.
The reason for the noisy result of latent traversal is the gaussian constrain which is not worked insufficiently. I seem the solution is increasing batch size and decreasing the learning rate. And also discriminator loss and total correlation loss is not balanced so it may be affected.
