Hi team, I'm using the AutoMFLES model but everytime I use it, it gets stuck on the last series. any idea on what is happening? Im running the code on databricks

Does it only happen with mfles?

That's correct

Is that series super long or have anything weird going on with it?

I don't think is that large, I have 952 unique_ids with max 167 samples

We recently changed the parallel processing, so that 800 is actually 800 partitions of the 952 ids and that last partition may have several series, but 13 minutes is a lot so there's most likely something going on there. So if you run that exact workflow with seasonal naive it completes successfully?

Yeah seasonalnaive runs in like 2sec

could be multiprocessing with LASSO @José Morales

we disable multithreading when n_jobs>1

ah ok nvm then.

Ok I change to polars and now I see this error:

ValueError: attempt to get argmin of an empty sequence

Can you include the full error?