Experiments for understanding disentanglement in VAE latent representations
-
Updated
Feb 2, 2023 - Python
Experiments for understanding disentanglement in VAE latent representations
Pytorch implementation of β-VAE
Pytorch implementation of FactorVAE proposed in Disentangling by Factorising(http://arxiv.org/abs/1802.05983)
Replicating "Understanding disentangling in β-VAE"
Deep active inference agents using Monte-Carlo methods
[ICML 2020] InfoGAN-CR and ModelCentrality: Self-supervised Model Training and Selection for Disentangling GANs
A TensorFlow implementation of FactorVAE, proposed in "Disentangling by Factorising" by Kim et al.
Generation of sprite dataset used when testing disentangled models.
To associate your repository with the dsprites topic, visit your repo's landing page and select "manage topics."