Created by: Termset

程序在一个强化学习环境（MaCA）的train/abc/下，在该环境的根目录下面执行程序。 environment/：推演环境 　　　interface.py：环境调用接口 agent/：用于存储封装好的可直接调用的智能体决策代码，每个决策代码可单独设立一个文件夹，按照标准封装，系统直接调用 　　　fix_rule/：基于规则的同构异构对抗决策算法 　　　fix_rule_no_att/：基于规则的同构异构对抗决策算法（无攻击能力） 　　　simple/：基于DQN的同构对抗决策示例算法，需使用pytorch obs_construct/：观测值组织形式自定义，开发者可根据算法需要自定义obs结构 　　　simple/：基于DQN的同构对抗决策示例对应的obs组织 fight_mp.py：对抗调度程序 replay.py 日志回放入口程序 maps/：对抗场景库 　　　1000_1000_2_10_vs_2_10.map：预置异构对抗场景 　　　1000_1000_fighter10v10.map：预置同构对抗场景 configuration/：配置数据 　　　reward.py：部分预置回报检查点的赋值 　　　system.py：部分系统推演规则定义 train/：决策算法训练所在路径 　　　simple/：simple决策模型训练代码

Actor部分程序如下：

from train.abc.maca_agent import MaCA_Agent
from train.abc.env_construct import env_construct

@parl.remote_class
class Actor(object):
    def __init__(self, config):
        self.config = config

        self.envs = []

        for _ in range(config['env_num']):
            # config['blue'] = blue_agent.get_obs_ind()
            env = Environment(config['map_path'], config['red_get_obs_ind'], config['blue_get_obs_ind'],
                              max_step=config['max_step'],
                              random_pos=config['random_pos'],
                              render=config['render'])
            env = env_construct(env)
            self.red_fighter_num = env.red_fighter_num
            env = MonitorEnv(env)
            self.envs.append(env)

单独运行Actor也是正常的，进行训练时会报错No module named 'train'

[07-12 22:19:50 MainThread @machine_info.py:86] nvidia-smi -L found gpu count: 1
[07-12 22:19:50 MainThread @machine_info.py:86] nvidia-smi -L found gpu count: 1
[07-12 22:19:50 MainThread @main.py:157] Waiting for 1 remote actors to connect.
[07-12 22:19:50 MainThread @main.py:161] Remote actor count: 1
Exception in thread Thread-4:
Traceback (most recent call last):
  File "/home/xtq/anaconda3/lib/python3.6/threading.py", line 916, in _bootstrap_inner
    self.run()
  File "/home/xtq/anaconda3/lib/python3.6/threading.py", line 864, in run
    self._target(*self._args, **self._kwargs)
  File "train/abc/main.py", line 172, in run_remote_sample
    remote_actor = Actor(self.config)
  File "/home/xtq/anaconda3/lib/python3.6/site-packages/parl/remote/remote_decorator.py", line 127, in __init__
    raise RemoteError('__init__', traceback_str)
parl.remote.exceptions.RemoteError: [PARL remote error when calling function `__init__`]:
No module named 'train'
traceback:
Traceback (most recent call last):
  File "/home/xtq/anaconda3/lib/python3.6/site-packages/parl/remote/job.py", line 298, in wait_for_connection
    cls = cloudpickle.loads(message[1])
ModuleNotFoundError: No module named 'train'

这个@parl.remote_class修饰的类，访问模块有什么特殊的吗？