https://github.com/stifler7/rl-for-game-ai

Game Artificial Intelligence with Reinforcement Learning This project provides experience in applying techniques of RL to game artificial intelligence. It shall demonstrate several RL algorithms- specifically Q-learning and deep Q-networks (DQN)- to game environments.
https://github.com/stifler7/rl-for-game-ai

deep-learning deep-neural-networks deep-reinforcement-learning machine-learning

Last synced: 12 days ago
JSON representation

Game Artificial Intelligence with Reinforcement Learning This project provides experience in applying techniques of RL to game artificial intelligence. It shall demonstrate several RL algorithms- specifically Q-learning and deep Q-networks (DQN)- to game environments.

• Host: GitHub
• URL: https://github.com/stifler7/rl-for-game-ai
• Owner: STiFLeR7
• Created: 2024-10-17T05:54:52.000Z (4 months ago)
• Default Branch: master
• Last Pushed: 2024-10-17T06:29:58.000Z (4 months ago)
• Last Synced: 2024-12-01T23:33:46.652Z (2 months ago)
• Topics: deep-learning, deep-neural-networks, deep-reinforcement-learning, machine-learning
• Language: Python
• Homepage: https://github.com/STiFLeR7/RL-for-Game-AI
• Size: 11.7 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Deep Q-Learning for CartPole Environment

This project implements a Deep Q-Learning agent to solve the CartPole environment using TensorFlow and PyTorch. The agent learns to balance a pole on a cart by taking actions based on the current state of the environment.

## Table of Contents

- [Installation](#installation)

- [Usage](#usage)

- [Training](#training)

- [Agent](#agent)

- [Dependencies](#dependencies)

## Installation

To set up the project, clone this repository and install the required dependencies:

```bash

git clone https://github.com/STiFLeR7/RL-for-Game-AI.git

cd RL-for-Game-AI

pip install -r requirements.txt

```

Make sure to install a compatible version of Gym and its dependencies.

## Usage

1. Ensure you have Python 3.7 or higher installed.

2. Activate your virtual environment (if you're using one).

3. Run the training script:

```bash

python train.py

```

4. The agent will train over a specified number of episodes and save the model as `dqn_model.pth`.

## Training

You can customize the number of training episodes in the `train_agent` function within the `train.py` script. The default is set to 1000 episodes.

## Agent

The agent uses a Deep Q-Network (DQN) approach to approximate the Q-values for each action based on the current state. It employs experience replay and epsilon-greedy strategy for action selection.

### DQN Architecture

- **Input Layer**: Takes the state representation from the environment.

- **Hidden Layers**: Fully connected layers with ReLU activations.

- **Output Layer**: Outputs Q-values for each possible action.

## Dependencies

The following Python packages are required to run this project:

- gym

- numpy

- torch

- tensorflow

- matplotlib

You can install them using pip:

```bash

pip install gym numpy torch tensorflow matplotlib

```

 project or raise issues if you encounter any problems.