PPO playing impala-QbertNoFrameskip-v4 from https://github.com/sgoodfriend/rl-algo-impls/tree/e47a44c4d891f48885af0b1605b30d19fc67b5af
35bb3c8
source benchmarks/train_loop.sh | |
# export WANDB_PROJECT_NAME="rl-algo-impls" | |
BENCHMARK_MAX_PROCS="${BENCHMARK_MAX_PROCS:-5}" | |
ALGOS=( | |
# "vpg" | |
# "dqn" | |
"ppo" | |
) | |
ENVS=( | |
"impala-PongNoFrameskip-v4" | |
"impala-BreakoutNoFrameskip-v4" | |
"impala-SpaceInvadersNoFrameskip-v4" | |
"impala-QbertNoFrameskip-v4" | |
"impala-CarRacing-v0" | |
) | |
train_loop "${ALGOS[*]}" "${ENVS[*]}" | xargs -I CMD -P $BENCHMARK_MAX_PROCS bash -c CMD | |