This message was deleted.

Most point prediction losses can work: MAE, MSE, sMAPE And other distribution prediction losses: MQL, DistributionLoss-Gaussian, Gaussian Mixture

Also we have multiple features as input

You can feed your features into the model through hist_exog, futr_exog and stat_exog lists

Yes but the model doesn't improve on Validation loss

Not all features need to have predictive power Some can also have redundant information that makes learning hard NeuralForecast does not have automated feature selection capabilities yet