PPO Agent Playing LunarLander-v3
This is a PPO agent trained on LunarLander-v3.
Evaluation
- Actual evaluation environment: LunarLander-v3
- Mean reward: -45.76
- Standard deviation: 15.41
- Evaluation episodes: 10
Algorithm
- Proximal Policy Optimization (PPO)
- Actor-Critic
- Generalized Advantage Estimation (GAE)
- Custom PyTorch implementation
Hugging Face Deep RL Course
Unit 8 Part I
Evaluation results
- mean_reward on LunarLander-v3self-reported-45.76 +/- 15.41