A collection of reinforcement learning projects exploring different algorithms and environments. Each project demonstrates a unique RL paradigm -- from on-policy actor-critic methods and policy gradients to imitation learning -- applied to challenging control and game-playing tasks.
| Project | Description | Algorithm | Environment |
|---|---|---|---|
| FIFA Actor-Critic | PPO agent that learns to play Google Football | Actor-Critic (PPO) | Google Football |
| Lunar Lander Policy Gradient | Policy gradient agent for LunarLander-v2 | Actor-Critic | OpenAI Gym LunarLander-v2 |
| Imitation Learning Cars | Self-driving via imitation learning using NVIDIA CNN | Behavioral Cloning | CARLA / Custom Dataset |
Each project is self-contained under projects/<project-name>/. To get up and running:
# 1. Clone the repository
git clone https://github.com/shivamakhauri04/reinforcement-learning.git
# 2. Navigate to the project of interest
cd reinforcement-learning/projects/<project-name>
# 3. Install dependencies
pip install -r requirements.txtTip: It is recommended to create a virtual environment (
venvorconda) before installing dependencies to avoid conflicts between projects.
Refer to each project's own README for detailed usage instructions, training commands, and evaluation steps.
reinforcement-learning/
├── README.md
├── LICENSE
├── .gitignore
├── .gitattributes
└── projects/
├── fifa-actor-critic/
├── lunar-lander-policy-gradient/
└── imitation-learning-cars/
Each project follows a standard layout:
projects/<project-name>/
├── README.md
├── requirements.txt
├── src/
├── checkpoints/
└── data/
This project is licensed under the MIT License -- see the LICENSE file for details.
Copyright (c) 2019 Shivam Akhauri