env = make_vec_env('LunarLander-v2', n_envs=24)
  model = PPO(
    policy = 'MlpPolicy',
    env = env,
    n_steps = 1024,
    batch_size = 64,
    n_epochs = 16,
    gamma = 0.999,
    gae_lambda = 0.98,
    ent_coef = 0.01,
    verbose=1)
  model.learn(total_timesteps=500000)