It looks like the models are not training, and I tried several approaches. Could other people share their experience?
My idea is that it might happen if labels do not correspond to the Field_IDs. It might happen if quality of the fields is not 3 = Good, but reversed. I am quite puzzled with the data. If I am not alone, is it possible for organizers to check the data, please?
What do you think?
seems, there was a problem with the code, it is better now.
nice score. Good work!