hi all i am following the nixtila mlforecast tutorial here and see that input dataset has multiple unique_id time series but only one model is trained. i had two questions • Is it possible to train separate models for each unique time series? • also does the library one handle this in Hierarchical Series example where we use ETS method which to my understanding can only be used for univariate time series? https://nixtlaverse.nixtla.io/mlforecast/index.html https://nixtlaverse.nixtla.io/hierarchicalforecast/examples/australiandomestictourism.html

mlforecast always trains a single global model. However, you can use sklearn models in statsforecast, which will train one per serie (guide)

got it thanks for clarifying 😄