PPO LunarLander-v2

Evaluation

Mean reward: 263.21 +/- 17.22

Score used by the Deep RL Course certification checker:

263.21 - 17.22 = 245.99

Important note

The PPO policy was originally trained using LunarLander-v3.

It was subsequently evaluated in the legacy LunarLander-v2 environment using Gymnasium 0.29.1 for compatibility with the legacy Hugging Face Deep RL Course certification checker.

The model therefore uses the same trained policy while the reported certification evaluation is a genuine evaluation on LunarLander-v2.

Model

Algorithm: PPO
Library: Stable-Baselines3
Environment evaluation: LunarLander-v2
Evaluation episodes: 10

Downloads last month: 10

Video Preview

Reinforcement Learning

Evaluation results

mean_reward on LunarLander-v2
self-reported

263.21 +/- 17.22