How can I get the 'peptide' and 'target' of train.csv when retrain the scorer

[Peggyying]

Hello,

I am interested in Spectralis. I have a few questions regarding retraining a random forest model regressor use spectralis.train_scorer_from_csvs.

How should I structure the training and validation datasets for retraining the model? What should be included in the 'peptide' and 'target' columns of the training set?
Is the 'peptide' column the peptide sequences predicted by the de novo model, or the true label sequences? And how to calculate the 'target' value?
Any guidance or example dataset you could provide would be greatly appreciated.

Thank you for your time and assistance.