DQN With PyTorch (CartPole)

Single-file PyTorch DQN example with training and evaluation on CartPole.

Run

1. Install dependencies:

   • pip install torch gymnasium numpy

2. Train and evaluate:

   • python "Neural Network by Pytorch.py"

Output

• Trains for 300 episodes by default.
• Saves model to dqn_cartpole.pt.
• Prints a final summary with rewards, loss, and eval score.