sgoodfriend's picture
A2C playing QbertNoFrameskip-v4 from https://github.com/sgoodfriend/rl-algo-impls/tree/0760ef7d52b17f30219a27c18ba52c8895025ae3
c41aee4
raw
history blame
217 Bytes
source benchmarks/train_loop.sh
ALGOS="ppo"
ENVS="PongNoFrameskip-v4 BreakoutNoFrameskip-v4"
BENCHMARK_MAX_PROCS="${BENCHMARK_MAX_PROCS:-3}"
train_loop $ALGOS "$ENVS" | xargs -I CMD -P $BENCHMARK_MAX_PROCS bash -c CMD