I reproduced it successfully. Attached is a screenshot. Expect the author to give you a Python script file for making your own data set.

[Wei-JL]

I trained normally and at about 120 rounds, the inference AUC value reached 96.35026351887952 % , a result that is very little different from the 97% mentioned in the paper. Thanks to the authors, can you please provide the script for making my own dataset?

[fanzz1208]

我正常训练，大约 120 轮，推断 AUC 值达到 96.35026351887952 %，这个结果与论文中提到的 97% 差别不大。感谢作者，你能提供制作我自己的数据集的脚本吗？

您还，请问可以将您复现成功的代码发给我吗 2563992699@qq.com

[HuYongting]

Excuse me, the model link given by the author has been invalid. Have you saved the pre-training model given by the author of the paper? Can you send me a copy? My email address is 2865613686@qq.com. Thank you very much.

[xiaoliukuaidianpao]

I trained normally and at about 120 rounds, the inference AUC value reached 96.35026351887952 % , a result that is very little different from the 97% mentioned in the paper. Thanks to the authors, can you please provide the script for making my own dataset?

Hello, my reproduction result is not ideal, can you send me a copy of your successful reproduction code? My E-mail address is liushimiao0304@163.com.