mirror of
https://github.com/vale981/ray
synced 2025-03-09 12:56:46 -04:00
42991d723f
Cleans up of the rllib/examples folder by moving all example Envs into rllibexamples/env (so they can be used by other scripts and tests as well).
45 lines
1.2 KiB
Python
45 lines
1.2 KiB
Python
import argparse
|
|
|
|
from ray.rllib.examples.env.stateless_cartpole import StatelessCartPole
|
|
|
|
parser = argparse.ArgumentParser()
|
|
parser.add_argument("--stop", type=int, default=200)
|
|
parser.add_argument("--torch", action="store_true")
|
|
parser.add_argument("--use-prev-action-reward", action="store_true")
|
|
parser.add_argument("--run", type=str, default="PPO")
|
|
parser.add_argument("--num-cpus", type=int, default=0)
|
|
|
|
if __name__ == "__main__":
|
|
import ray
|
|
from ray import tune
|
|
|
|
args = parser.parse_args()
|
|
|
|
ray.init(num_cpus=args.num_cpus or None)
|
|
|
|
configs = {
|
|
"PPO": {
|
|
"num_sgd_iter": 5,
|
|
"vf_share_layers": True,
|
|
"vf_loss_coeff": 0.0001,
|
|
},
|
|
"IMPALA": {
|
|
"num_workers": 2,
|
|
"num_gpus": 0,
|
|
"vf_loss_coeff": 0.01,
|
|
},
|
|
}
|
|
|
|
tune.run(
|
|
args.run,
|
|
stop={"episode_reward_mean": args.stop},
|
|
config=dict(
|
|
configs[args.run], **{
|
|
"env": StatelessCartPole,
|
|
"model": {
|
|
"use_lstm": True,
|
|
"lstm_use_prev_action_reward": args.use_prev_action_reward,
|
|
},
|
|
"use_pytorch": args.torch,
|
|
}),
|
|
)
|