Dennis Bader

Hey, for the given `n` steps you want to predict, you have to provide more data for your future covariates (should extend more into the future). We explain the minimum...

Hi @kpiaskowski and thanks for writing. Our current `load_model()` method doesn't support direct mapping to another torch device. It loads the model in the same state as it was saved....

Hey @Insiyaa, you can use `model.historical_forecasts()` for that. Here is a link to the docs: https://unit8co.github.io/darts/generated_api/darts.models.forecasting.arima.html#darts.models.forecasting.arima.ARIMA.historical_forecasts

Have a look at [MissingValuesFiller](https://unit8co.github.io/darts/generated_api/darts.dataprocessing.transformers.missing_values_filler.html#darts.dataprocessing.transformers.missing_values_filler.MissingValuesFiller) :)

Do they share the same time index? The metrics extract the overlapping part (non-empty) of the target and prediction indices in order to compute the metric. If below line raises...

Hi @albin02t , the `metric` must be a callable accepting 2 TimeSeries as input arguments and that returns a float value. Darts' mase has 3 positional input arguments including `insample`...

You can with any other metrics that accepts two poisitonal input arguments (see all metrics [here](https://unit8co.github.io/darts/generated_api/darts.metrics.metrics.html), like rmse, mse, mae, marre, ..) Edit: as a heads up, multiple time series...

Hi @gdevos010 and thanks for another PR. Tests are failing when you upgrade PyTorch Lightning and PyTorch to the newest versions. I'm working on a fix in #1124. Could you...

Hi @matanost-netop and thanks for reporting. What you're saying is right, if you use TensorBoard logging and manually save/load the model to/from a different directory, the trainer can‘t find the...

I thought that we fixed this issue in darts version 0.19.0. Could it be that you are saving in a different environment than you are loading the model?