BDA test results

Hi, thank you for the providing the code!
I have trained the BDA model on my own device using the same settings as described in your paper. However, the harmonic mean I obtained is around 0.765, which is slightly lower than the results reported in the paper.
Is this difference normal? Could you please provide your pre-trained weight files?

Best regards