Official PyTorch implementation for MMVAE+, introduced in the paper "MMVAE+: Enhancing the Generative Quality of Multimodal VAEs without Compromises", published at ICLR 2023.

We thank the authors for the code MMVAE from which our codebase is based.

mkdir data 
cd data 
curl -L -o data_ICLR.zip https://polybox.ethz.ch/index.php/s/Eg4RVLC0XxnKAJX/download
unzip data_ICLR.zip 
unzip PolyMNIST.zip