diff --git a/examples/alphazero/README.md b/examples/alphazero/README.md index 9f5dbeb25..f23b6cc1e 100644 --- a/examples/alphazero/README.md +++ b/examples/alphazero/README.md @@ -8,7 +8,7 @@ Note that you need to install `jax` and `jaxlib` in addition to the packages wri ```sh $ pip install -r requirements.txt -$ python3 train.py env_id=go_9x9 selfplay_batch_size=1024 seed=0 +$ python3 train.py env_id=go_9x9 seed=0 ``` ## Reference @@ -20,5 +20,6 @@ chess, shogi, and go through self-play" ## Change history +- **[#1088](https://github.com/sotetsuk/pgx/pull/1088)** Adjust to API v2 ([wandb report](https://api.wandb.ai/links/sotetsuk/0g44pjsg)). - **[#1055](https://github.com/sotetsuk/pgx/pull/1055)** Use default Gumbel AlphaZero hyperparameters ([wandb report](https://api.wandb.ai/links/sotetsuk/o8752t54)). - **[#1026](https://github.com/sotetsuk/pgx/pull/1026)** Initial version. Supposed to reproduce the Pgx paper results ([wandb report](https://api.wandb.ai/links/sotetsuk/5q30e5n9)).