Ivan

runopp does not converge for case5

I couldn't reproduce your problem. I tried a small dataframe, the DeepAR model from the sample and your settings, and got a sequence of scalar predictions. Do you mind testing...

I'm assuming `filtered_base_data` does not have the "total_sales" column. Since you are using the same parameters as `training` (in which `target` is set to "total_sales") to create `inference_ts_dataset`, it complains...

@vishnu020 if that solved your problem, consider closing the issue.

I'm currently using pytorch-forecasting 1.0.0 and have the same problem when trying to pickle a model like `TemporalFusionTransformer`. The problem seems to be that one of its super classes [`TupleOutputMixIn`](https://pytorch-forecasting.readthedocs.io/en/stable/api/pytorch_forecasting.utils.TupleOutputMixIn.html)...

If you do this, then you are predicting the entirety of the testing set all at once, not in a sliding window fashion.

See https://pytorch.org/get-started/locally/. You install PyTorch by `pip install torch`.

You can refer to [this example](https://pytorch-forecasting.readthedocs.io/en/stable/tutorials/stallion.html). See Cell 5 where `training`and `validation` are defined.

`Cython.Compiler.Errors.CompileError: sklearn\svm\_liblinear.pyx` Looks like a problem with installation of [sklearn](https://github.com/scikit-learn/scikit-learn) actually.