ray/rllib/tuned_examples/es/cartpole-es.yaml

12 lines
282 B
YAML

cartpole-es:
env: CartPole-v0
run: ES
stop:
episode_reward_mean: 150
timesteps_total: 500000
config:
# Works for both torch and tf.
use_pytorch: false
num_workers: 2
noise_size: 25000000
episodes_per_batch: 50