the 9th formulation was inconsistence with your open code

[eeric]

code: self.t = target_logit.mean() 0.01 + (1 - 0.01) self.t paper: t(k) = αr(k) + (1 − α)t(k−1) α = 0.99, r(k) = sum cosθyi arget_logit.mean() in your code was not same with sum cosθyi in your paper