zeynpakn/ppo-LunarLander-v2

🚀 LunarLander-v2 — PPO Agent

A Deep Reinforcement Learning agent trained using Proximal Policy Optimization (PPO) with Stable-Baselines3.

⚙️ Training

• Environment: LunarLander-v2
• Algorithm: PPO (MlpPolicy)
• Training: 1,000,000 timesteps
• Parallel environments: 16
• Framework: Gymnasium & Stable-Baselines3

📊 Results

• Mean Reward: 250.74
• Standard Deviation: 23.30
• Evaluation Score (Mean − Std): 227.44
• Hugging Face Recorded Score: 219.19
• Unit 1: Passed ✅

🔗 Project & Source Code

Explore the GitHub repository for the training notebook, implementation, and gameplay demo. Part of the Hugging Face Deep RL Course — Unit 1.

Houston, we had a problem. Then PPO learned how to land. 🌙🚀