diff --git a/README.md b/README.md index cfce1d1f..6689eac5 100644 --- a/README.md +++ b/README.md @@ -231,9 +231,9 @@ python -m pax.experiment +train_iters=1e3 +num_envs=10 ``` We currently support two major environments: `MatrixGames` and `CoinGame`. -``` For `MatrixGames`, we support the ability to specify your own payoff matrix either through the CLI or the `yaml` files. For example the common Iterated Prisoners Dilemma is: + ```bash python -m pax.experiment +experiment/ipd=ppo ++payoff="[[-2,-2], [0,-3], [-3,0], [-1,-1]]" ++wandb.group="testing" ```