Hi, I really enjoy Neuralforecast! How can I forecast just one timestep in the future with an offset from the input sequence? As in TensorFlow tutorial https://www.tensorflow.org/tutorials/structured_data/time_series

Hi @Oleksii Kaplenko! Not right now,

neuralforecast

does not have a weighting value per future timestamp, you will need to forecast the complete window after the input sequence until the desired timestamp

Cool thanks! adding covariate variables support for all transformers would be also very appreciated