May I ask you what baseline you use for your comparison?

[roomo7time]

Hi, I'd like to ask you some questions on the baseline model that you used for the comparison.

In README.md, the reported baseline gives a simialr performace to that of the proposed RPL. May I ask you what you exactly mean by this 'baseline'?

Is it the one trained with conventional softmax cross entropy? Then, what score did you use in the inference stage (the identification stage)?

Thanks!