How can I train this scoring function on my own dataset?

[Dadiao-shuai]

Dear developer, I truly want to train this RF-Score-VS with my own datasets to do further research, but I could find insructions in this Github page. It would be nice if you could offer some information, thanks!