dqn_lunar_v2 / results.json
exploiter345's picture
DQN lunar lander V2 trained for 500k, n_steps=2048, batch_size=128
fd9048b
{"mean_reward": 167.08045521990226, "std_reward": 79.19141170577636, "is_deterministic": true, "n_eval_episodes": 10, "eval_datetime": "2022-05-09T00:27:45.321964"}