Exercies from Sleep stage classification from polysomnography (PSG) data Page

The task is to: Fetch 50 subjects from the Physionet database and run a 5-fold cross-validation leaving each time 10 subjects out in the test set.

Do you have a suggestion on how to do this?

To use the eeg_power_band function ([epochs (mne.Epochs — MNE 1.2.0 documentation)), it only accepts mne epoch data. Hence, I concatenate the data for 40 subjects in the training set and 10 subjects in the testing set to run pipe = make_pipeline(FunctionTransformer(eeg_power_band, validate=False), RandomForestClassifier(n_estimators=100, random_state=42)).

Could I create an array of mne epoch to run instead? How would I do this?