RL-LunarLander-PPo

PPO Agent playing LunarLander-v2

This is a trained model of a PPO agent playing LunarLander-v2 using the stable-baselines3 library.

Usage (with Stable-baselines3)

This is the Lunar Lander PPO solution with 1000000 total_timesteps with a mean reward of 268 +/- 20.50

from huggingface_sb3 import load_from_hub, package_to_hub
from huggingface_hub import (
    notebook_login,
) 

from stable_baselines3 import PPO
from stable_baselines3.common.env_util import make_vec_env
from stable_baselines3.common.evaluation import evaluate_policy
from stable_baselines3.common.monitor import Monitor

env = make_vec_env('LunarLander-v2', n_envs=16)
model = PPO(
    policy='MlpPolicy',
    env=env,
    n_steps=1024,
    batch_size=32,
    n_epochs=4,
    gamma=0.999,
    gae_lambda=0.98,
    ent_coef=0.01,
    verbose=1
)
model.learn(total_timesteps=1000000, progress_bar=True)
model_name = "ppo-lunarlander-v2"
model.save(model_name)
eval_env = Monitor(gym.make('LunarLander-v2'))
mean_reward, std_reward = evaluate_policy(model, eval_env, n_eval_episodes=10, deterministic=True)
print(f'Mean reward = {mean_reward:.2f} +/- {std_reward:.2f}')
...