A2C playing CartPole-v1 from https://github.com/sgoodfriend/rl-algo-impls/tree/0760ef7d52b17f30219a27c18ba52c8895025ae3
464d478
| source benchmarks/train_loop.sh | |
| ALGOS="ppo" | |
| ENVS="CartPole-v1 MountainCar-v0 MountainCarContinuous-v0 Acrobot-v1 LunarLander-v2" | |
| BENCHMARK_MAX_PROCS="${BENCHMARK_MAX_PROCS:-3}" | |
| train_loop $ALGOS "$ENVS" | xargs -I CMD -P $BENCHMARK_MAX_PROCS bash -c CMD | |