Let's use est._model instead of est._models[-1], in case FFX decides on a more sophisticated method of choosing a model from the Pareto front in future.
Hmm, not sure if I'm doing this PR right. Github seems to have included in the PR the merge I did from your branch to my fork. Sorry, I'm a Github amateur!
Let's use est._model instead of est._models[-1], in case FFX decides on a more sophisticated method of choosing a model from the Pareto front in future.