PPO playing MicrortsAttackShapedReward-v1 from https://github.com/sgoodfriend/rl-algo-impls/tree/587a52bc38901314c7c1b5c6892acf9315796cf3
a1df559
while getopts a:e:s:p: flag | |
do | |
case "${flag}" in | |
a) algo=${OPTARG};; | |
e) envs=${OPTARG};; | |
s) seeds=${OPTARG};; | |
p) project_name=${OPTARG};; | |
esac | |
done | |
WANDB_TAGS=$(bash scripts/tags_benchmark.sh) | |
project_name="${project_name:-rl-algo-impls-benchmarks}" | |
seeds="${seeds:-1 2 3}" | |
for env in $(echo $envs); do | |
for seed in $seeds; do | |
echo python train.py --algo $algo --env $env --seed $seed --pool-size 1 --wandb-tags $WANDB_TAGS --wandb-project-name $project_name --virtual-display | |
done | |
done | |