DQN

Deep-Q-Learning

This repo is an implementation of deep-q-learning. It demonstrates the agent learning to play atari breakout using pytorch

Requirements:

1. Python3
2. Gym
3. Pytorch

Usage:

• To train the agent:

  python main.py --train_dqn

• To test the agent:

  python main.py --test_dqn --model_test_path "trained_model_path"

Results:

• In Training:

• Running the agent for 100 episodes

  • Average reward: 64