samuelscheit / carcassonne-ai Goto Github PK

I tried testing the models that you've trained in src/results/carcassonne2/ using the testing code in muzero.py (running the script, picking 1 to load the model and then 3 to self play) and I'm getting some strange results.

In the screenshot below, you can see that many of the moves being made are illegal. I suspect that this is either an issue with the way that legal actions are defined in https://github.com/Flam3rboy/carcassonne-ai/blob/a3a96cc0650538d78bc60e78ecf73b6e013b279d/src/muzero/carcassonne/utils/action_util.py#L43-L76
Alternatively, it might be that the interpretation of selected actions is off.

I'm happy to help fix the problem, but on first glance, I didn't quite understand how ActionUtil works. Do you have any ideas here? It's also entirely possible that I'm using the code in the wrong way or that I've misunderstood something else, so let me know if that's the case!

This error happened when I try "3. Render some self play games"
I try a few day, but I can't fix it.

Exception has occurred: RayTaskError(AttributeError)
�[36mray::SelfPlay.play_game()�[39m (pid=13172, ip=127.0.0.1, repr=<self_play.SelfPlay object at 0x0000027D88204C70>)
File "python\ray_raylet.pyx", line 663, in ray._raylet.execute_task
File "python\ray_raylet.pyx", line 667, in ray._raylet.execute_task
File "python\ray_raylet.pyx", line 614, in ray._raylet.execute_task.function_executor
File "C:\Users\s0708.conda\envs\muzero\lib\site-packages\ray_private\function_manager.py", line 701, in actor_method_executor
return method(__ray_actor, *args, **kwargs)
File "C:\Users\s0708.conda\envs\muzero\lib\site-packages\ray\util\tracing\tracing_helper.py", line 462, in _resume_span
return method(self, *_args, **_kwargs)
File "C:\Users\s0708\Desktop\carcassonne-ai-main\src\muzero\self_play.py", line 124, in play_game
self.game.render()
File "C:\Users\s0708\Desktop\carcassonne-ai-main\src\muzero\carcassonne2.py", line 218, in render
self.env.render()
File "C:\Users\s0708\Desktop\carcassonne-ai-main\src\muzero\carcassonne\carcassonne_game.py", line 60, in render
self.visualiser.draw_game_state(self.state)
AttributeError: 'NoneType' object has no attribute 'draw_game_state'
File "C:\Users\s0708\Desktop\carcassonne-ai-main\src\muzero\muzero.py", line 402, in test
ray.get(
File "C:\Users\s0708\Desktop\carcassonne-ai-main\src\muzero\muzero.py", line 677, in
muzero.test(render=True, opponent="human", muzero_player=0)
ray.exceptions.RayTaskError(AttributeError): �[36mray::SelfPlay.play_game()�[39m (pid=13172, ip=127.0.0.1, repr=<self_play.SelfPlay object at 0x0000027D88204C70>)
File "python\ray_raylet.pyx", line 663, in ray._raylet.execute_task
File "python\ray_raylet.pyx", line 667, in ray._raylet.execute_task
File "python\ray_raylet.pyx", line 614, in ray._raylet.execute_task.function_executor
File "C:\Users\s0708.conda\envs\muzero\lib\site-packages\ray_private\function_manager.py", line 701, in actor_method_executor
return method(__ray_actor, *args, **kwargs)
File "C:\Users\s0708.conda\envs\muzero\lib\site-packages\ray\util\tracing\tracing_helper.py", line 462, in _resume_span
return method(self, *_args, **_kwargs)
File "C:\Users\s0708\Desktop\carcassonne-ai-main\src\muzero\self_play.py", line 124, in play_game
self.game.render()
File "C:\Users\s0708\Desktop\carcassonne-ai-main\src\muzero\carcassonne2.py", line 218, in render
self.env.render()
File "C:\Users\s0708\Desktop\carcassonne-ai-main\src\muzero\carcassonne\carcassonne_game.py", line 60, in render
self.visualiser.draw_game_state(self.state)
AttributeError: 'NoneType' object has no attribute 'draw_game_state'

Is this code below cause this problem?
In carcassonne2.py

class Game(AbstractGame):
    """
    Game wrapper.
    """
    def __init__(self, seed=None):
        self.env = CarcassonneGame(players=2,
                                   tile_sets=[TileSet.BASE],
                                   supplementary_rules=[],
                                   visualiser=None)

My enviroment is

python=3.9.17

numpy=1.25.2
torch=2.0.1
tensorboard=2.10.0
gym=0.26.2
ray=1.12.0
seaborn=0.12.2
nevergrad=0.12.0
redis=5.0.0
graphviz=0.20.1

May you share your current requirement? Thanks a lot. 🙏🙏