Validation on target test set

[adricarda]

Thank you for releasing the code.

I was wondering whether you used the target dataset for selecting the final checkpoint. In the paper you say that you pick the best model according to the source domain (as it should be). Hovever, in train_SPST.py it seems to me that you validate on the target domain? Is my understanding correct?

Thank you.

[zou-longkun]

Thanks for the reminder, the code has been updated