-
Notifications
You must be signed in to change notification settings - Fork 1
[Class] SupervisingTrainer 📦
João Saraiva edited this page Jun 9, 2022
·
1 revision
A Supervising Trainer is a Pipeline Unit. Meaning, you can instantiate one and include it in a Pipeline.
Give a (1) Supervised Model and a list (1...*) of Supervised Train Conditions. You may also name the Supervising Trainer.
model = SupervisedModel(...)
conditions1 = SupervisedTrainConditions(...)
conditions2 = SupervisedTrainConditions(...)
conditions3 = SupervisedTrainConditions(...)
(...)
trainer = SupervisingTrainer(model, (conditions1, conditions2, conditions3, ...))
Add it to a Pipeline:
pipeline.add(trainer)
Apply it directly by calling the method apply
. You need to give:
-
1...* object: A collection of Timeseries that are equally equally segmented. These are to be fed as input to the model during trains and tests.
-
1 target: A Timeseries to be compared to the model output during trains and tests.
trainer.apply(object = (timeseries1, timeseries2, ...), target = target_timeseries)
The trainer will iterate through every set of conditions defined upon instantiation, and report performance metrics and plots for each.
Comming soon.