Hi Nixtla team ! I want to perform custom training with weight on sample. Everything works fine for training. However, when I want to perform cross validation with the same logic, I don't see any easy way to do that. Especially to create CV windows, then apply preprocessing and do a custom fit on each model. Am I missing something ?

Hey. We don't have anything for that at the moment, however we have some utilities for generating the splits which can help. That function returns a generator of cutoffs, train, valid, so you can iterate through those and perform your current training in each fold

Alright, wanted to be sure that there was no other way to do that. Thanks for your insights !

Feel free to open an issue for this. Training with weights is something that's been requested a couple of times and I think we could add it