Open mayii2001 opened 4 months ago
Same problem here. Did you solve it? thanks!
Same problem here. There should be a problem with the way the library is written in metrics, which is not maintained by the authors.
The same error occurred, and I solved it by simply applying the loss function:
mean_losses = SMAPE(reduction="none")
mean_losses = mean_losses.loss(predictions.output, predictions.y)
Same error here. Not sure how it works but here are my codes
mean_losses = SMAPE(reduction="none")
mean_losses = mean_losses.loss(predictions.output, predictions.y[0]).mean(1)
Since it seems like the predictions.y is a tuple, and the shape of the predictions.y[0] is the same with the predictions.output
Expected behavior
I executed code in Tutorial of Demand forecasting with the Temporal Fusion Transformer and expected to get result of SMAPE. The issue appeared while i didn`t change any code from tutorial. I hit the breakpoint but the breakpoint can't break before the error. Solving this problem is beyond my power.
Code to reproduce the problem