Multimodal Variational Autoencoder Implementation of "Multimodal Generative Models for Scalable Weakly-Supervised Learning" paper Requirements torch == 0.4.1 torchvision == 0.2.1 Usage python3 train.py