Hello, team.
Thanks for your wonderful model!
This issue existed in the closed list, but when i ran the model to try the forecasting task, i met the same problem too.
And follow your advice in that issue to reload the model, it still can't be fixed.
My tool is Pycharm+python3.11, with cuda11.8+torch2.0.1.
As you said in 'forecasting.ipynb', i loaded the model and use data of ETTh1.csv, the debug process stopped on the line of 'loss = criterion(output.forecast, forecast)', Error Message showed ”AttributeError: 'NoneType' object has no attribute 'size'“， and the final error line was 'if not (target.size() == input.size()):' in 'functional.py'.
After checking, the main reason maybe on the line of 'output = model(x_enc=timeseries, input_mask=input_mask)', beacuse the Forecast of output was None.
Beg your reply!
Hello, team.
Thanks for your wonderful model!
This issue existed in the closed list, but when i ran the model to try the forecasting task, i met the same problem too.
And follow your advice in that issue to reload the model, it still can't be fixed.
My tool is Pycharm+python3.11, with cuda11.8+torch2.0.1.
As you said in 'forecasting.ipynb', i loaded the model and use data of ETTh1.csv, the debug process stopped on the line of 'loss = criterion(output.forecast, forecast)', Error Message showed ”AttributeError: 'NoneType' object has no attribute 'size'“， and the final error line was 'if not (target.size() == input.size()):' in 'functional.py'.
After checking, the main reason maybe on the line of 'output = model(x_enc=timeseries, input_mask=input_mask)', beacuse the Forecast of output was None.
Beg your reply!