Views
No views yet
1model = PPO('MlpPolicy', env, n_steps = 1024, batch_size = 64, n_epochs = 4, gamma = 0.999, gae_lambda = 0.98, ent_coef = 0.01, verbose=1)
2
3model.learn(total_timesteps=1000000)
4
5model_name = "ppo-LunarLander-v2"
6model.save(model_name)
7
8
9eval_env = Monitor(gym.make("LunarLander-v2"))
10
11# Evaluate the model with 10 evaluation episodes and deterministic=True
12mean_reward, std_reward = evaluate_policy(model, eval_env, n_eval_episodes=10, deterministic=True)
13
14print(f"mean_reward={mean_reward:.2f} +/- {std_reward}")
15
16# mean_reward=266.997 +/- 21.582
17