How to predict large unseen tabular datas with trained model?

I have had the same issue, here is an elegant solution :slight_smile: