Learning the Prediction Distribution for Semi-Supervised Learning with Normalising Flows

This codebase contains PyTorch implementation of the paper:

Learning the Prediction Distribution for Semi-Supervised Learning with Normalising Flows. Ivana Balažević*, Carl Allen*, and Timothy Hospedales. arXiv:2007.02745. [Paper]

Wide ResNet code taken from https://github.com/szagoruyko/wide-residual-networks/tree/master/pytorch.

Normalising flow code taken from https://github.com/karpathy/pytorch-normalizing-flows for the classification experiments and https://github.com/bayesiains/nflows for the image-to-image experiments.

Image-to-image translation code coming soon...