Exception raised in creation task: The actor died because of an error raised in its creation task, ray::RolloutWorker.__init__()

[chatsmile]

What happened + What you expected to happen

A simple example but fails TAT Exception raised in creation task: The actor died because of an error raised in its creation task, ray::RolloutWorker.init()

Failure # 1 (occurred at 2024-06-28_21-09-46) The actor died because of an error raised in its creation task, ray::PPO.init() (pid=227821, ip=192.168.0.110, actor_id=baeb6937fabac97b336e66f101000000, repr=PPO) File "/home/yujintong/anaconda3/envs/torch/lib/python3.8/site-packages/ray/rllib/evaluation/worker_set.py", line 229, in _setup self.add_workers( File "/home/yujintong/anaconda3/envs/torch/lib/python3.8/site-packages/ray/rllib/evaluation/worker_set.py", line 682, in add_workers raise result.get() File "/home/yujintong/anaconda3/envs/torch/lib/python3.8/site-packages/ray/rllib/utils/actor_manager.py", line 497, in _fetch_result result = ray.get(r) ray.exceptions.RayActorError: The actor died because of an error raised in its creation task, ray::RolloutWorker.init() (pid=227899, ip=192.168.0.110, actor_id=2800cc7cdc700cfb962c6dba01000000, repr=<ray.rllib.evaluation.rollout_worker.RolloutWorker object at 0x7f198c773b80>) File "/home/yujintong/anaconda3/envs/torch/lib/python3.8/site-packages/ray/rllib/evaluation/rollout_worker.py", line 407, in init self.env = env_creator(copy.deepcopy(self.env_context)) File "/home/yujintong/anaconda3/envs/torch/lib/python3.8/site-packages/ray/rllib/env/utils/init.py", line 125, in _gym_env_creator env = env_descriptor(env_context) TypeError: init() takes 1 positional argument but 2 were given

During handling of the above exception, another exception occurred:

ray::PPO.init() (pid=227821, ip=192.168.0.110, actor_id=baeb6937fabac97b336e66f101000000, repr=PPO) File "/home/yujintong/anaconda3/envs/torch/lib/python3.8/site-packages/ray/rllib/algorithms/algorithm.py", line 533, in init super().init( File "/home/yujintong/anaconda3/envs/torch/lib/python3.8/site-packages/ray/tune/trainable/trainable.py", line 161, in init self.setup(copy.deepcopy(self.config)) File "/home/yujintong/anaconda3/envs/torch/lib/python3.8/site-packages/ray/rllib/algorithms/algorithm.py", line 631, in setup self.workers = WorkerSet( File "/home/yujintong/anaconda3/envs/torch/lib/python3.8/site-packages/ray/rllib/evaluation/worker_set.py", line 181, in init raise e.args[0].args[2] TypeError: init() takes 1 positional argument but 2 were given

Versions / Dependencies

Ray 2.10.0 python 3.8

Code below

from future import absolute_import from future import division from future import print_function

import argparse import gymnasium as gym

from Task_env import ASEnv import ray from ray.tune import register_env from ray.rllib.models import ModelCatalog from ray.rllib.algorithms.ppo import PPOConfig from ray.rllib.env.multi_agent_env import ENV_STATE from ray import air from ray import tune from ray import train from ray.rllib.algorithms.algorithm import Algorithm

if name == "main": parser = argparse.ArgumentParser() parser.add_argument("--num-iters", type=int, default=1000) parser.add_argument("--num-workers", type=int, default=16) args = parser.parse_args()

ray.init()
def policy_mapping_fn(agent_id):
    return agent_id

config = PPOConfig()
config.environment(env=ASEnv)
config.training(train_batch_size=8000)
config.resources(num_gpus=0)
config.rollouts(num_rollout_workers=args.num_workers, sample_async=False)
config.multi_agent(
    policies={
        agent: (None, gym.spaces.Box(-1.0, 1.0, (18,)),
                gym.spaces.MultiDiscrete([3,3]), {
                    "gamma": 0.99,
                    "model": {
                        "use_lstm": True
                    }
                }) for agent in ["Agent{}".format(i) for i in range(3)] #  + ["main"]
    },
    policy_mapping_fn=policy_mapping_fn
)
#训练

tune.Tuner(
    "PPO",
    run_config=air.RunConfig(
        checkpoint_config=train.CheckpointConfig(checkpoint_frequency=1, num_to_keep=5,
                                                 checkpoint_score_attribute='episode_reward_mean'),
        name="IPPO",
        stop={
            "timesteps_total": 40000,
        }
    ),
    param_space=config.to_dict(),
).fit()

Issue Severity

High: It blocks me from completing my task.

[simonsays1980]

@chatsmile Thanks for raising this issue. I tried to execute your example but I am missing some modules you are using in there. Could you provide a reproducable example please?

Also your version is very old. Is it possible for you to lift up onto our new stack and give Ray 2.31 a try?

[chatsmile]

这是来自QQ邮箱的假期自动回复邮件。   您好，我最近正在休假中，无法亲自回复您的邮件。我将在假期结束后，尽快给您回复。