[Question] do we have acess to training scripts

[Mohamed209]

Dear THUDM MobileRL Team,

I am a researcher working on efficient RL training for mobile GUI agents.
Your MobileRL paper is excellent work, and I would like to build upon it
for my research on parameter-efficient training.

I noticed the GitHub repo has evaluation code but not training code.
Would you be willing to share the training implementation (AdaGRPO,
AdaPR, SPA) for academic research purposes? I will cite your paper
and can share my findings on model compression with you.

Thank you for your time and contribution to the community.