jchen8tw / ewn-gym

EinStein würfelt nicht gymnasium environment.

ewn-gym

EinStein würfelt nicht for gymnasium

How to train a Stable-baseline-3 model ?

setup the environment

git clone https://github.com/jchen8tw/ewn-gym
cd ewn-gym
pip install -r requirements.txt

python train.py -op [opponent policy] {A2C,PPO}
# For more options
python train.py -h

How to evaluate a SB3 model

python eval_A2C.py --opponent_policy [path to model or policy name] --model [path to your model]

How to train an Alpha Zero agent

conda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidia
git clone https://github.com/jchen8tw/alpha-zero-general
cd alpha-zero-general
pip install -r requirements.txt
python main_ewn.py

How to eval all policies

Move your alpha zero models for temp into alpha_zero_models folder with corresponding board_size(e.g 5x5, 7x7 .. etc) or pull it from the huggingface repo using git submodule.

cd alpha_zero_models
git submodule init
git submodule update

Now you can eval all policies except for the stable-baseline-3 policy(we've abandoned it for lack performance) .

python eval_pairs.py