-
Notifications
You must be signed in to change notification settings - Fork 552
Optimizing over a list defining points in the search space? #371
Comments
The short answer is no. The longer answer is that you can have categorical dimensions, which means only the values in that list will be tried, however the optimizer will visit all possible combinations. For example if you have two categorical dimensions: If that includes too many points you don't want to visit, then right now the best option would be to inherit from If there is demand or interest we could add such a class to scikit-optimize directly, so contributions welcome. |
Ok, thank you. How exactly though is a |
You can't ... sorry for this. In my head I was convinced that you could pass a I think it would be a great feature to have and shouldn't be too tricky to implement. |
This is not completely fixed by #373 right? |
#373 only made it possible to pass around an instance of |
Is there a feature where, instead of using dimensions to define the search space, skopt will optimize on a list of points which define the search space?
This would be a useful feature if I wanted to optimize in a subset of the possible points of a space (rather than the entire space).
The text was updated successfully, but these errors were encountered: