Is the trainer.gen_algo the finally trained model on which i can continue training? I would like to clone a handcrafted controller into a stable-baselines agent and than continue training. It seems...