5使用spinningup的DDPG算法时出现问题，求老师帮忙看看

第十课训练机械臂时，把PPO算法换成DDPG后，一直出现warning：Warning: could not pickle state_dict. 我其他参数都没有改变，代码截图如下，已经研究好几天无法解决了，求老师帮忙看看！！！万分感谢！！！

from spinup import ppo_pytorch as ppo
from spinup import ddpg_pytorch as ddpg
from UR5_Controller import UR5_Controller
from spinup.utils.test_policy import load_policy_and_env, run_policy
import torch

TRAIN = 1
target = [0, -1.57, 1.57, 0, 0, 0]
env = lambda : UR5_Controller(target)

if TRAIN:
    ac_kwargs = dict(hidden_sizes=[128,128,128], activation=torch.nn.ReLU)
    logger_kwargs = dict(output_dir='log', exp_name='ur5_goToTarget')

    ddpg(env, ac_kwargs=ac_kwargs, logger_kwargs=logger_kwargs,
        steps_per_epoch=2000, epochs=4000)

else:
    _, get_action = load_policy_and_env('log')
    env_test = UR5_Controller(target)
    run_policy(env_test, get_action)