Reputation: 1825
Is there any documentation where I could details regarding this kind of stuff? For example:
import gym
# environment for agent
env = gym.make('Pendulum-v0')
env.reset()
print(env.observation_space.high, env.observation_space.low)
# pendulum observation space ranges from [-1, -1, -8] to [1, 1, 8]
I cant figure out what each number in observation space means. I guess two of them are x and y coordinates (although I dont know which is which), and what does the third number stand for?
Upvotes: 2
Views: 2688
Reputation: 59
Open AI's documentation covers this:
Dimensionality of Observation space, Action space, reward, etc. of all the environments: https://github.com/openai/gym/wiki/Table-of-environments
Specific ranges for them, e.g. https://github.com/openai/gym/wiki/Pendulum-v1
In your case, assuming Pendulum-v0 shares v1's features, your three values mean:
Best
Upvotes: 0
Reputation: 1825
Apparently openai
's github page has a wiki which contains the information I was looking for: https://github.com/openai/gym/wiki
Upvotes: 1