metadata
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: 274.81 +/- 20.36
name: mean_reward
verified: false
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)
TODO: Add your code
from huggingface_hub import notebook_login
from stable_baselines3 import PPO
from stable_baselines3.common.evaluation import evaluate_policy
from stable_baselines3.common.env_util import make_vec_env
# Create the environment
env = make_vec_env('LunarLander-v2', n_envs=16)
model = PPO(
policy = 'MlpPolicy', # The policy to be optimized
env = env, # The environment in which the agent will act
n_steps = 2048, # The number of steps to run for each environment per update
batch_size = 64, # Minibatch size
n_epochs = 10, # Number of epoch when optimizing the surrogate loss
gamma = 0.999, # discount factor used to weigh future rewards in the total reward calculation
gae_lambda = 0.98, # parameter used in the Generalized Advantage Estimation (GAE) algorithm
ent_coef = 0.01, # Entropy coefficient for the loss calculation
verbose=0) # Verbosity level: 0 for no output, 1 for info messages (such as device or wrappers used), 2 for debug messages
# Train it for 1,500,000 timesteps
model.learn(total_timesteps=1500000, progress_bar=True)
# Specify file name for model and save the model to file
model_name = "ppo-LunarLander-v2"
model.save(model_name)
# Create a new environment for evaluation
eval_env = gym.make("LunarLander-v2")
# Evaluate the model with 10 evaluation episodes and deterministic=True
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}")