Personalized Federal learning(PFL) allows clients to cooperatively train a personalized model without disclosing their private dataset. However, PFL suffers from Non-IID, Heterogeneous Devices, lack of fairness, and unclear contribution which urgently need the →