HaoYin
HaoYin
https://github.com/deezer/sigir2019-2stagesampling/blob/ab9de4e520cd6da703d7309ffd22f7a4638fd7cd/samplers/spreadout_sampler.py#L66 This calculation is inconsistent with the original formula. in the paper: ```math (1-s^2)^{\frac{d-1}{2} - 1} ``` while in the code: ```math (1-s^2)^{1- \frac{d-1}{2}} ``` Is it a bug or...