You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I was trying to delete some of the solutions of the population on the fly (inside the fitness_func) which were introducing my prediction to diverge. This is possible only when the flag parallel_processing is not set, while when working with multiple processes any changes in the attributes of the class will not be reflected in the subsequent generations.
I would like to know if it is possible to extend this feature to parallel processing as well