Accompanying code for research paper examining whether similarity partitioning for n-fold cross-validation during hyper-parameter optimisation can help build models more capable of OOD generalisation. - View it on GitHub