Open
Description
I found that the parameters set by the author for different network models are different, which is different from the parameter Settings used in TimeXer training. Shouldn't the model parameters set in the training process be the same?
I am confused about the different training results caused by the different parameters.
Metadata
Metadata
Assignees
Labels
No labels