is this speaker embedding better than X-vector?

Hi, Thank you for sharing this amazing package!

I want to compare this model's perf with other well known speaker embedding models. (i.e. X-vector)
Before running experiments myself, I wonder whether there are already published performance comparisons (i.e. EER in speaker verification)

Any comment would be much appreciated!