Towards end-to-end likelihood-free inference with convolutional neural networks (Radev, Mertens, Voss, & Köthe)

This repository contains the Jupyter notebooks for each of the three examples presented in the paper. There are two additional folders named img and model_checkpoints with the latter containing a pretrained network (trained_model.hdf5)

The datasets used for the training of the models can be downloaded from the following link:

https://heidata.uni-heidelberg.de/privateurl.xhtml?token=fd1753a6-363a-4815-8605-a3dc03268797

There is one zip-file for each example, containing the training set and the test set. The Jupyter Notebooks assume you have the data in a folder named data on the same level as the two folders model_checkpoints and img.

Please make sure to unzip the content of the respective zip-file into the data folder you created.