This message was deleted.

Hi @허성민! 1. Can you try running our examples on exogenous variables: https://nixtla.github.io/neuralforecast/examples/exogenous_variables.html. There might be an issue with your data or the way you are passing the variables. 2. Yes, adding more training iterations might help to converge. Remember adding a validation set (

val_size>0

) and early stopping (

early_stop_patience_steps>0

) to prevent overfitting. 3. Yes, its more than enough! We have examples where

NHITS

produce SOTA results with a lot less records (see the Getting Started with AirPassengers for example).

I received your answer well. As you mentioned, I tried again using the sample code and was able to get the result normally. Thank you.