Save Results

With the class persist options you can specify different locations where the result log of the hyperparameter search can be stored. You can save them into a folder into a MongoDB (you can have both or you leave out one of the parameters). You also have the possibility to choose if you want to save all the predictions (Beware - may be infinitely large) and/or to store the feature importances of your models.

from photonai.base.PhotonBase import Hyperpipe, OutputSettings

o_settings = OutputSettings(mongodb_connect_url="mongodb://localhost:27017/photon_db",
                                save_predictions=False,
                                save_feature_importances=False,
                                project_folder="/home/usr1/proj23/")


my_pipe = Hyperpipe('perfect_pipe',
                    ...,
                    output_settings=o_settings)

Save and load the best model

Of course you can save and share your model. Here is how: