PPO LunarLander-v2 — Unit 8
This repository contains a Proximal Policy Optimization (PPO) agent trained from scratch using PyTorch for the LunarLander-v2 environment, as part of the Hugging Face Deep Reinforcement Learning Course, Unit 8.
Training details
- Algorithm: Proximal Policy Optimization (PPO)
- Environment: LunarLander-v2
- Framework: PyTorch
- Training timesteps: 500,000
- Model file:
ppo_lunarlander.pt
Evaluation
The agent was evaluated over 10 episodes. The observed mean reward was -4.62, with a standard deviation of 40.46. These are results from the local evaluation and may vary between runs.
Files
ppo_lunarlander.pt: Trained model weights.