search

facebookresearch / hanabi_SAD

Simplified Action Decoder for Deep Multi-Agent Reinforcement Learning
Other
96 stars 35 forks source link

How can I generate the Figure 5 of the Other-Play paper { https://arxiv.org/pdf/2003.02979.pdf }? #25

Open VigRAN084 opened 3 years ago

VigRAN084 commented 3 years ago

Hello, How can I generate the Figure 5 of the Other-Play paper { https://arxiv.org/pdf/2003.02979.pdf }?

Thanks, Vignesh Rangarajan

VigRAN084 commented 3 years ago

Hi Hengyuan,

Thank you for letting me know. I went ahead and created a new issue on the GitHub { https://github.com/facebookresearch/hanabi_SAD/issues/25 }.

Appreciate your help.

Best, Vignesh Rangarajan

On Jul 21, 2021, at 12:28 PM, Hengyuan Hu @.***> wrote:

Hello,

I created a new function in the hanabi_env.h to generate the list of legal moves performed by the agents. Here is the function declaration:

const std::vector getLegalMoves() const{ return state_->MoveHistory(); } When the code is run, I get an error that error: ‘HanabiHistoryItem’ was not declared in this scope.

I also added a .def("get_legal_moves", &HanabiEnv::legal_moves) in the pybind.cc http://pybind.cc/ file.

Could you let me know how I can resolve this error? Appreciate if you could let me know.

Thanks, Vignesh Rangarajan

hengyuan-hu commented 3 years ago

It is not included in the repo. I will try to find the script add it here later.

VigRAN084 commented 3 years ago

Thanks! @hengyuan-hu

hengyuan-hu commented 3 years ago

checkout the tools/action_matrix.py from the latest master.