Can I use cross-learning with Chronos2 using the dataframe API?

[sebassaras02]

Hi everyone

I started to implement chronos2 in my forecasting activities. My question is more related to the fact if I can apply cross learning during the predict_df method.

I really like the dataframe API 'cos it reduces the overhead of data wrangling and dates generation. From my POV, I've tried some experiment providing directly the argument of predict_batches_jointly in the predict_df method, and the results are different.

Is this result fine? Can I use in this way? Should I use using as kwargs? Is only available cross-learning for the tensors method?

I will thank your answer.

Kind regards

Sebastian Sarasti

[shchur]

Yes, you can pass this kwarg to predict_df and it will be used under the hood.