FileNotFoundError:

[Throne-myworld]

when I download the code and run arnold.py using pycharm in Windows, things go like that:

FileNotFoundError: [Errno 2] No such file or directory: 'D:\PyCharmproject\Arnold-master\Arnold-master\dumped\default\vhzihaaocq\train.log'
Can this programm run successfully using Win10?

[markcutajar]

I think the package was designed for linux which allows automatic creation of folders vhzihaaocq for example. However, Windows doesn't allow that to get it to work, changes need to be done to pre-create the folder than use the log.

Checking it out, in src/utils.py all the

subprocess.Popen("mkdir %s" % main_dump_path, shell=True).wait()

lines or similar needs to be changed to

os.makedirs(_path_name)

[Throne-myworld]

I think the package was designed for linux which allows automatic creation of folders vhzihaaocq for example. However, Windows doesn't allow that to get it to work, changes need to be done to pre-create the folder than use the log.

Checking it out, in src/utils.py all the

subprocess.Popen("mkdir %s" % main_dump_path, shell=True).wait()

lines or similar needs to be changed to

os.makedirs(_path_name)

good work, my friend!

[Throne-myworld]

@markcutajar ，but a new issue comes,:
File "D:\PyCharmproject\Arnold-master\Arnold-master\src\doom\game.py", line 133, in init
assert os.path.isfile(self.scenario_path)
AssertionError
when I try to print the self.scenario_path, it says:D:\PyCharmproject\Arnold-master\Arnold-master\resources\scenarios\ .wad, there is no filename before'.wad', it really confused me.
Also, the inti_py in the model folder is empty, so how should I do to run the code?
using arnold.py or init_py?

[markcutajar]

I don't think that error is related, to run the script make sure you run it as explained in the readme:

python arnold.py --scenario defend_the_center --action_combinations "turn_lr+attack" --frame_skip 2

I think the issue is that if you just run python.arnold.py the scenario is not given and it is not captured earlier in the processes. Better error handling is definitely needed here.

[Throne-myworld]

Better error handling is definitely needed here.

OK , I will try it as soon as possible.

[Throne-myworld]

@markcutajar I tried it , the new issue says
File "D:\PyCharmproject\Arnold-master\Arnold-master\src\model\dqn\base.py", line 81, in base_forward
output = torch.cat([conv_output] + embeddings, dim=1)
RuntimeError: torch.cat(): Tensors must have same number of dimensions: got 2 and 3

[Trenza1ore]

@markcutajar I tried it , the new issue says File "D:\PyCharmproject\Arnold-master\Arnold-master\src\model\dqn\base.py", line 81, in base_forward output = torch.cat([conv_output] + embeddings, dim=1) RuntimeError: torch.cat(): Tensors must have same number of dimensions: got 2 and 3

This is one of the issues related to old PyTorch (0.3.x in this case) vs newer version, I have made a PyTorch 2.x compatible fork that should work as intended on modern version of various first/third party libraries (note that I have not tested NumPy 2.x although at a glance it should work).