AI
Lunar Lander · DQN
Reinforcement-learning agent that learns to land a lander.
- Python
- PyTorch
Details
- Context
- Epitech · 2026
- Role
- Implementation, training, evaluation, documentation
- Method
- Discrete DQN (PyTorch, Gymnasium) · compared with random and heuristic agents · replay-buffer ablation · wind variant
Images
Trained DQN agent Reward per episode during training