sgoodfriend's picture
PPO playing impala-BreakoutNoFrameskip-v4 from https://github.com/sgoodfriend/rl-algo-impls/tree/e47a44c4d891f48885af0b1605b30d19fc67b5af
b05c680
raw
history blame
381 Bytes
source benchmarks/train_loop.sh
# export WANDB_PROJECT_NAME="rl-algo-impls"
BENCHMARK_MAX_PROCS="${BENCHMARK_MAX_PROCS:-3}"
ALGOS=(
# "vpg"
# "dqn"
"ppo"
)
ENVS=(
"procgen-coinrun-easy"
"procgen-starpilot-easy"
"procgen-bossfight-easy"
"procgen-bigfish-easy"
)
train_loop "${ALGOS[*]}" "${ENVS[*]}" | xargs -I CMD -P $BENCHMARK_MAX_PROCS bash -c CMD