BrokenPipeError: [Errno 32] Broken pipe

I have created a tensortrade.env.generic.environment.TradingEnv using tensortrade.env.default.create as follows:

env_train = default.create(
    action_scheme=action_scheme,
    reward_scheme=reward_scheme,
    portfolio=portfolio_train,
    feed=feed_train,
    window_size=10,
    renderer="screen-log",
    enable_logger=True
)

The above environment works fine as it is but my model: PPO utilizes the GPU only around 0-3 %. Therefore, I wanna create a vector environment consisting of multiple environments for multiprocessing, and this is how I do it:

from stable_baselines3.common.env_util import make_vec_env
vec_env_train = make_vec_env(env_train, n_envs=num_envs, seed=42, vec_env_cls=SubprocVecEnv)

That results in BrokenPipeError: [Errno 32] Broken pipe. The error has been discussed in 509 and according to that the above code should work.

Clarification: The error relates to SB3, not TensorTrade. The following is the complete trace of the error:

---------------------------------------------------------------------------
BrokenPipeError                           Traceback (most recent call last)
<ipython-input-13-3b5897de4e37> in <module>()
      3 from stable_baselines3.common.env_util import make_vec_env
      4 
----> 5 vec_env_train = make_vec_env(env, n_envs=4, seed=42, vec_env_cls=SubprocVecEnv)

4 frames
/usr/local/lib/python3.7/dist-packages/stable_baselines3/common/env_util.py in make_vec_env(env_id, n_envs, seed, start_index, monitor_dir, wrapper_class, env_kwargs, vec_env_cls, vec_env_kwargs, monitor_kwargs, wrapper_kwargs)
    103         vec_env_cls = DummyVecEnv
    104 
--> 105     return vec_env_cls([make_env(i + start_index) for i in range(n_envs)], **vec_env_kwargs)
    106 
    107 

/usr/local/lib/python3.7/dist-packages/stable_baselines3/common/vec_env/subproc_vec_env.py in __init__(self, env_fns, start_method)
    108             work_remote.close()
    109 
--> 110         self.remotes[0].send(("get_spaces", None))
    111         observation_space, action_space = self.remotes[0].recv()
    112         VecEnv.__init__(self, len(env_fns), observation_space, action_space)

/usr/lib/python3.7/multiprocessing/connection.py in send(self, obj)
    204         self._check_closed()
    205         self._check_writable()
--> 206         self._send_bytes(_ForkingPickler.dumps(obj))
    207 
    208     def recv_bytes(self, maxlength=None):

/usr/lib/python3.7/multiprocessing/connection.py in _send_bytes(self, buf)
    402             # Also note we want to avoid sending a 0-length buffer separately,
    403             # to avoid "broken pipe" errors if the other end closed the pipe.
--> 404             self._send(header + buf)
    405 
    406     def _recv_bytes(self, maxsize=None):

/usr/lib/python3.7/multiprocessing/connection.py in _send(self, buf, write)
    366         remaining = len(buf)
    367         while True:
--> 368             n = write(self._handle, buf)
    369             remaining -= n
    370             if remaining == 0:

BrokenPipeError: [Errno 32] Broken pipe

### Checklist

• I have read the documentation (required)
• I have checked that there is no similar issue in the repo (required)
• I have checked my env using the env checker (required)
• I have provided a minimal working example to reproduce the bug (required)

[zeeshanalipanhwar]

It turns out the make_vec_env function takes env_id to be either str or Callable, not a gym.env as mentioned in its function details. The following code resolves the issue.

from stable_baselines3.common.env_util import make_vec_env
vec_env_train = make_vec_env(lambda:env_train, n_envs=num_envs, seed=42, vec_env_cls=SubprocVecEnv)

[araffin]

The following code resolves the issue.

yes, although you should do that with a lambda… it will give the same env object instead of instantiating new ones for each env.