Skip to main content

Prabhasak's group workspace

Timestamps visible
2020-09-25 08:00:19
Episode length: 1000.00 +/- 0.00
2020-09-25 08:31:51
Eval num_timesteps=798720, episode_reward=3430.46 +/- 2.58
2020-09-25 08:31:51
Episode length: 1000.00 +/- 0.00
2020-09-25 08:59:29
Eval num_timesteps=899072, episode_reward=3374.80 +/- 5.99
2020-09-25 08:59:29
Episode length: 1000.00 +/- 0.00
2020-09-25 09:33:47
Eval num_timesteps=999424, episode_reward=3306.72 +/- 4.02
2020-09-25 09:33:47
Episode length: 1000.00 +/- 0.00
2020-09-25 09:34:14
Choosing best saved model throughout training, instead of model available at end of training (EvalCallback enabled!)
2020-09-25 09:34:18
Trained Hopper-v2 with gail. Some stats:
2020-09-25 09:34:18

2020-09-25 09:40:00
[100]/100 successful episodes
2020-09-25 09:40:00

2020-09-25 09:40:00
Mean return:  3597.791
2020-09-25 09:40:00
Std return:  2.8599203
2020-09-25 09:40:00
Max return:  [3605.425]
2020-09-25 09:40:00
Min return:  [3590.562]
2020-09-25 09:40:00
Mean episode len:  1000.0
2020-09-25 09:40:00
Optimal policy found!
2020-09-25 09:40:00