| --- |
| library_name: stable-baselines3 |
| tags: |
| - LunarLander-v2 |
| - deep-reinforcement-learning |
| - reinforcement-learning |
| - stable-baselines3 |
| model-index: |
| - name: PPO |
| results: |
| - task: |
| type: reinforcement-learning |
| name: reinforcement-learning |
| dataset: |
| name: LunarLander-v2 |
| type: LunarLander-v2 |
| metrics: |
| - type: mean_reward |
| value: 241.85 +/- 48.02 |
| name: mean_reward |
| verified: false |
| --- |
| |
| # **PPO** Agent playing **LunarLander-v2** |
| This is a trained model of a **PPO** agent playing **LunarLander-v2**, trained for 1e6 time steps, obtaining: |
| **mean_reward** = 241.85 +/- 48.02 |
| |
| using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3). |
| |
| ## Usage (with Stable-baselines3) |
| |
| |
| ```python |
| import gym |
| |
| from stable_baselines3 import PPO # Modelo que vamos a usar |
| from stable_baselines3.common.evaluation import evaluate_policy # Evaluación de los resultados del modelo entrenado |
| from stable_baselines3.common.env_util import make_vec_env |
| |
| # Creo el env |
| env = gym.make('LunarLander-v2') |
| |
| # Selecciono el modelo, en este caso el PPO, y lo ponemos a entrenar |
| model = PPO('MlpPolicy',env,verbose=1).learn(total_timesteps=1000000,progress_bar=True) |
| |
| # Lo guardamos |
| model.save('Lunar_Lander') |
| |
| # Creamos un nuevo env en el que probamos el modelo (valdría el mismo pero reseteado) |
| eval_env = gym.make('LunarLander-v2') |
| |
| # Evaluamos el modelo |
| mean_reward, std_reward = evaluate_policy(model, eval_env, n_eval_episodes=10, deterministic=True) |
| |
| # Print the results |
| print(f"mean_reward={mean_reward:.2f} +/- {std_reward}") |
| |
| ``` |
