Run cmd python -m qlib.rl.contrib.train_onpolicy --config_path exp_configs/train_opds.yml --run_backtest with error

python -m qlib.rl.contrib.train_onpolicy --config_path exp_configs/train_opds.yml --run_backtest

/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/site-packages/tianshou/env/venvs.py:66: UserWarning: You provided an environment generator that returned an OpenAI Gym environment. We strongly recommend transitioning to Gymnasium environments. Tianshou is automatically wrapping your environments in a compatibility layer, which could potentially cause issues.
warnings.warn(
Traceback (most recent call last):
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/runpy.py", line 194, in _run_module_as_main
return _run_code(code, main_globals, None,
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/runpy.py", line 87, in _run_code
exec(code, run_globals)
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/site-packages/qlib/rl/contrib/train_onpolicy.py", line 268, in
main(config, run_training=not args.no_training, run_backtest=args.run_backtest)
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/site-packages/qlib/rl/contrib/train_onpolicy.py", line 241, in main
train_and_test(
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/site-packages/qlib/rl/contrib/train_onpolicy.py", line 159, in train_and_test
train(
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/site-packages/qlib/rl/trainer/api.py", line 63, in train
trainer.fit(vessel)
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/site-packages/qlib/rl/trainer/trainer.py", line 224, in fit
self.vessel.train(vector_env)
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/site-packages/qlib/rl/trainer/vessel.py", line 171, in train
collector = Collector(self.policy, vector_env, VectorReplayBuffer(self.buffer_size, len(vector_env)))
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/site-packages/tianshou/data/collector.py", line 80, in init
self.reset(False)
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/site-packages/tianshou/data/collector.py", line 131, in reset
self.reset_env(gym_reset_kwargs)
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/site-packages/tianshou/data/collector.py", line 147, in reset_env
obs, info = self.env.reset(**gym_reset_kwargs)
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/site-packages/qlib/rl/utils/finite_env.py", line 233, in reset
for i, o in zip(request_id, super().reset(request_id)):
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/site-packages/tianshou/env/venvs.py", line 280, in reset
assert (
AssertionError: The environment does not adhere to the Gymnasium's API.
Exception ignored in: <function DataQueue.del at 0x7fc212bf9dc0>
Traceback (most recent call last):
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/site-packages/qlib/rl/utils/data_queue.py", line 148, in del
self.cleanup()
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/site-packages/qlib/rl/utils/data_queue.py", line 101, in cleanup
self._queue.get(block=False)
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/multiprocessing/queues.py", line 111, in get
res = self._recv_bytes()
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/multiprocessing/connection.py", line 216, in recv_bytes
buf = self._recv_bytes(maxlength)
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/multiprocessing/connection.py", line 414, in _recv_bytes
buf = self._recv(4)
File "/hdd9/yanzheng/folder_to_make_space_for_home/miniconda3/envs/py38/lib/python3.8/multiprocessing/connection.py", line 383, in _recv
raise EOFError
EOFError:

Reproduce the command with exp_configs/train_opds.yml and --run_backtest, then trace the reset path from qlib/rl/contrib/train_onpolicy.py through qlib/rl/trainer/vessel.py and qlib/rl/utils/finite_env.py into Tianshou's venvs.py. Check the environment reset contract and determine what needs to change so the command passes the Gymnasium API assertion without the cleanup error.