Fixed a bug in the RCIA-LVQ implementation, in which the filter was always random. Additionally cleaned up the helpers file.

Added more samples to the Red MNIST dataset, so that the performance is more reliable. Added new jobs for the Red MNIST dataset.

Added the original VisTex images used to generate the VisTex dataset. These images consists of the classes Food, Fabric, Brick and Bark.

Added the wrangling/cleaning scripts used to clean the dataset and produce the final data used in the paper.

Renamed the parameter lamb to reg to better reflect its purpose. Also removed regularization for the prototypes, since this is rather pointless.