This is the second release of the information maximising neural network code used in the paper "Principled physical inference with neural networks.". It is more efficient and general than v1.0 although works in the same way.
v.1.0
This is the first release of the information maximising neural network code used in the paper "Principled physical inference with neural networks." It contains all the machinery to create the data in the paper, and can be used to rerun any analysis in the paper. The network itself is complete for use with any external data, although the protocols for use must be of the form of the data created in the paper.