JAndretti / Simple_DDQN_Cartpole

Simple_DDQN_Cartpole: Applying Double Deep Q-Networks to train an agent for balancing a pole on a cart. Reinforcement learning project focuses on efficient learning and improved decision-making for mastering the Cartpole control problem.

Simple_DDQN_Cartpole

Simple_DDQN_Cartpole: Applying Double Deep Q-Networks to train an agent for balancing a pole on a cart. Reinforcement learning project focuses on efficient learning and improved decision-making for mastering the Cartpole control problem.

Files

• agent.py: Contains classes for the DDQN agent and its training algorithm.
• train.py: Initializes the Cartpole environment and trains the agent. It saves model weights and displays training curves.
• test.py: Uses saved model weights to test the agent's inference performance.

Usage

1. Install the necessary dependencies: Python 3.x, PyTorch, NumPy.
2. Run train.py to train the DDQN agent on the Cartpole environment.
3. After training, model weights are saved for future use.
4. Run test.py to evaluate the trained agent's performance using the saved weights.
5. Training progress and inference results can be monitored through the displayed curves and outputs.