Can we make running more parallelizable?

[kudkudak]

• Add id_fold parameter to fit_active_learning.
• Modify run_experiment_grid to call folds and return GridSearchExperimen
• Add new function run_experiments based on run_experiment_grid

[kudkudak]

Yes we can, will be refactored before PLOS strategy - I have a neat plan that will make code really nice and readable