Step function while creating custom environments

RLlib
1

I was browsing the documentation for creating environments in RLlib

class MyEnv(gym.Env):
def init(self, env_config):
self.action_space = <gym.Space>
self.observation_space = <gym.Space>
def reset(self):
return
def step(self, action):
return , <reward: float>, <done: bool>, <info: dict>

Should the step function return the done signal or the terminated and truncated signals?

2

It should return terminated and truncated; could you link where you found this? We’ll update it accordingly.

1 Like
3

Thank you for your prompt reply. I found it here: Environments — Ray 2.3.1

4

Opening a PR to fix this along with some other minor doc fixes. Thanks for raising this!