Hi all,
I am attempting to configure my own MultiAgentEnv, using ray[rllib]==1.6.0. I get the following error when I try to train with a PPOTrainer:
/opt/conda/lib/python3.7/site-packages/ray/_private/services.py:238: UserWarning: Not all Ray Dashboard dependencies were found. To use the dashboard please install Ray using `pip install ray[default]`. To disable this message, set RAY_DISABLE_IMPORT_WARNING env var to '1'.
warnings.warn(warning_message)
Loading environment football failed: No module named 'gfootball'
2021-09-15 23:17:13,132 INFO trainer.py:714 -- Tip: set framework=tfe or the --eager flag to enable TensorFlow eager execution
2021-09-15 23:17:13,132 INFO ppo.py:159 -- In multi-agent mode, policies will be optimized sequentially by the multi-GPU optimizer. Consider setting simple_optimizer=True if this doesn't work for you.
2021-09-15 23:17:13,133 INFO trainer.py:728 -- Current log_level is WARN. For more information, set 'log_level': 'INFO' / 'DEBUG' or use the -v and -vv flags.
(pid=1462) Loading environment football failed: No module named 'gfootball'
Traceback (most recent call last):
File "/kaggle/working/LuxPrivate/training.py", line 65, in <module>
trainer = ppo.PPOTrainer(env=LuxEnv, config=config)
File "/opt/conda/lib/python3.7/site-packages/ray/rllib/agents/trainer_template.py", line 136, in __init__
Trainer.__init__(self, config, env, logger_creator)
File "/opt/conda/lib/python3.7/site-packages/ray/rllib/agents/trainer.py", line 592, in __init__
super().__init__(config, logger_creator)
File "/opt/conda/lib/python3.7/site-packages/ray/tune/trainable.py", line 103, in __init__
self.setup(copy.deepcopy(self.config))
File "/opt/conda/lib/python3.7/site-packages/ray/rllib/agents/trainer_template.py", line 146, in setup
super().setup(config)
File "/opt/conda/lib/python3.7/site-packages/ray/rllib/agents/trainer.py", line 739, in setup
self._init(self.config, self.env_creator)
File "/opt/conda/lib/python3.7/site-packages/ray/rllib/agents/trainer_template.py", line 175, in _init
num_workers=self.config["num_workers"])
File "/opt/conda/lib/python3.7/site-packages/ray/rllib/agents/trainer.py", line 827, in _make_workers
logdir=self.logdir)
File "/opt/conda/lib/python3.7/site-packages/ray/rllib/evaluation/worker_set.py", line 85, in __init__
lambda p, pid: (pid, p.observation_space, p.action_space)))
File "/opt/conda/lib/python3.7/site-packages/ray/_private/client_mode_hook.py", line 82, in wrapper
return func(*args, **kwargs)
File "/opt/conda/lib/python3.7/site-packages/ray/worker.py", line 1623, in get
raise value
File "python/ray/_raylet.pyx", line 640, in ray._raylet.task_execution_handler
File "python/ray/_raylet.pyx", line 488, in ray._raylet.execute_task
File "python/ray/_raylet.pyx", line 604, in ray._raylet.execute_task
ray.exceptions.RayActorError: The actor died because of an error raised in its creation task, ray::RolloutWorker.__init__() (pid=1462, ip=172.19.2.2)
File "/opt/conda/lib/python3.7/site-packages/ray/rllib/evaluation/rollout_worker.py", line 401, in __init__
self.env = env_creator(env_context)
File "/opt/conda/lib/python3.7/site-packages/ray/rllib/agents/trainer.py", line 1647, in <lambda>
register_env(name, lambda config: env_object(config))
TypeError: __init__() missing 1 required positional argument: 'debug'
(pid=1463) 2021-09-15 23:17:45,760 ERROR worker.py:428 -- Exception raised in creation task: The actor died because of an error raised in its creation task, ray::RolloutWorker.__init__() (pid=1463, ip=172.19.2.2)
(pid=1463) File "/opt/conda/lib/python3.7/site-packages/ray/rllib/evaluation/rollout_worker.py", line 401, in __init__
(pid=1463) self.env = env_creator(env_context)
(pid=1463) File "/opt/conda/lib/python3.7/site-packages/ray/rllib/agents/trainer.py", line 1647, in <lambda>
(pid=1463) register_env(name, lambda config: env_object(config))
(pid=1463) TypeError: __init__() missing 1 required positional argument: 'debug'
(pid=1463) Loading environment football failed: No module named 'gfootball'
(pid=1462) 2021-09-15 23:17:45,734 ERROR worker.py:428 -- Exception raised in creation task: The actor died because of an error raised in its creation task, ray::RolloutWorker.__init__() (pid=1462, ip=172.19.2.2)
(pid=1462) File "/opt/conda/lib/python3.7/site-packages/ray/rllib/evaluation/rollout_worker.py", line 401, in __init__
(pid=1462) self.env = env_creator(env_context)
(pid=1462) File "/opt/conda/lib/python3.7/site-packages/ray/rllib/agents/trainer.py", line 1647, in <lambda>
(pid=1462) register_env(name, lambda config: env_object(config))
(pid=1462) TypeError: __init__() missing 1 required positional argument: 'debug'
The training script:
import sys
sys.path.append('/kaggle/working/LuxPrivate')
from my_multilux_interface import MyInterface
import ray
ray.init()
from ray.tune.registry import register_env
from multilux.lux_env import LuxEnv
def env_creator(env_config):
# I pass env_config as {} and set defaults here instead
configuration = env_config.get('configuration', {})
debugg = env_config.get('debugggg', False)
interface = env_config.get('interface', MyInterface)
agents = env_config.get('agents', (None, "simple_agent"))
return LuxEnv(configuration, debugg,
interface=interface,
agents=agents)
register_env("multilux", env_creator)
# (3) Instantiate agent ------------------------------------------------------
import random
from ray.rllib.agents import ppo
config = {
"env_config": {},
"multiagent": {
"policies": {
# the first tuple value is None -> uses default policy
"unit-1": (None,
MyInterface.obs_spaces["unit"],
MyInterface.act_spaces["unit"],
{"gamma": 0.85}),
"unit-2": (None,
MyInterface.obs_spaces["unit"],
MyInterface.act_spaces["unit"],
{"gamma": 0.99}),
"citytile": (None,
MyInterface.obs_spaces["citytile"],
MyInterface.act_spaces["citytile"],
{}),
},
"policy_mapping_fn":
lambda agent_id:
"citytile" # Citytiles always have the same policy
if agent_id.startswith("ct_")
else random.choice(["unit-1", "unit-2"]) # Randomly choose from unit policies
},
}
trainer = ppo.PPOTrainer(env=LuxEnv, config=config) # <---------------- ERROR HERE
# (4) Train away -------------------------------------------------------------
while True:
print(trainer.train())
Any help or pointers much appreciated!
I knew that! Thank you so much!!