Look into difficulty of wrapping a Model as a scikit-learn Estimator (see: https://scikit-learn.org/stable/developers/develop.html). If not too difficult, this would allow for using NEMS Models directly with a lot of useful scikit tools like sklearn.model_selection.cross_val_score without needing to re-implement everything from scratch.
Look into difficulty of wrapping a Model as a scikit-learn Estimator (see: https://scikit-learn.org/stable/developers/develop.html). If not too difficult, this would allow for using NEMS Models directly with a lot of useful scikit tools like
sklearn.model_selection.cross_val_scorewithout needing to re-implement everything from scratch.