RFR: Continual Few-shot Reinforcement Learning