This project is created based on Playing Atari with deep RL, as a reproduction of Playing Atari with Deep Reiforcement Learing
DQN
├── model.py
├── models
│ ├── Boxing
│ │ ├── # saved model weights checkpoints for Boxing
│ ├── Breakout
│ │ ├── # saved model weights checkpoints for Breakout
│ └── Pong
│ ├── # saved model weights checkpoints for Pong
├── params.py
├── player.py
├── __pycache__
│ ├── # complied results
├── replay.py
├── trainer.py
├── videos
│ ├── Boxing
│ │ ├── # videos of agent playing Boxing games
│ ├── Breakout
│ │ ├── # videos of agent playing Breakout games
│ └── Pong
│ ├── # videos of agent playing Pong games
└── wrappers.py
If you want to work on this repository or train your own models, you can simply change the ModelTrainer(env_name=['your_atari_name']) in trainer.py and get the result of agent playing game with ModelPlayer in player.py