jnforja/ppo-LunarLander-v2-cleanrl

PPO Agent Playing LunarLander-v2

This model was trained locally with a CleanRL-style single-file PPO implementation for Hugging Face Deep RL Course Unit 8.

Results

• Mean reward: -180.47 +/- 74.10
• Evaluation episodes: 10
• Timesteps: 50000

Files

• model.pt: PyTorch policy checkpoint.
• results.json: evaluation results.
• replay.mp4: rendered policy preview.
• logs/: TensorBoard logs from the training run.