Rendering Problem on Windows 10 about gym-anm

[robinhenry]

@diegofz I believe this is due to a known issue when using python multiprocessing on Windows (e.g., see here).

gym-anm uses multiprocessing to render the environment in your browser: a local server is started in the background and communicates with the main thread using websockets to update the visualization in real-time.

In order to get the rendering to work with your setup (you shouldn't have this problem if you remove the env.render() statement), all you need to do is to guard the entry point of your code with an if __name__ == '__main__' statement.

For example, you can rewrite the example given in the Quickstart section:

import gym
import time

env = gym.make('gym_anm:ANM6Easy-v0')
o = env.reset()

for i in range(100):
    a = env.action_space.sample()
    o, r, done, info = env.step(a)
    env.render()
    time.sleep(0.5)  # otherwise the rendering is too fast for the human eye.
env.close()

as follows:

import gym
import time

def run():
    env = gym.make('gym_anm:ANM6Easy-v0')
    o = env.reset()
    
    for i in range(100):
        a = env.action_space.sample()
        o, r, done, info = env.step(a)
        env.render()
        time.sleep(0.5)  # otherwise the rendering is too fast for the human eye.
    env.close()

if __name__ == '__main__':
    run()

and that should do the trick. It worked for me on my Windows virtual machine.

Unfortunately I'm not sure there's a simple solution to get that to work in Jupyter Notebooks or in Collab. I would recommend you don't render the environment if you're using either of these tools.

Hope that helps!

[sifatron]

I have been having rendering issues, where nothing shows up, just a blank screen.
Based on this closed issue, I switched from windows 10 to windows 11. Also switched from Jupyter notebook to pycharm. And I am running the code you shared. I am still getting a blank screen, How can I fix it? Both machines have decent processors and gpu capabilities.