https://github.com/timothewt/deeprl
Implementation of some Deep Reinforcement Learning algorithms and environments.
https://github.com/timothewt/deeprl
a2c action-masking deep-reinforcement-learning dqn gym multiagent multiagent-reinforcement-learning pettingzoo ppo reinforcement-learning torch
Last synced: 2 months ago
JSON representation
Implementation of some Deep Reinforcement Learning algorithms and environments.
- Host: GitHub
- URL: https://github.com/timothewt/deeprl
- Owner: timothewt
- License: mit
- Created: 2023-10-03T14:25:54.000Z (over 1 year ago)
- Default Branch: master
- Last Pushed: 2023-10-26T13:18:38.000Z (over 1 year ago)
- Last Synced: 2025-01-18T18:41:55.373Z (4 months ago)
- Topics: a2c, action-masking, deep-reinforcement-learning, dqn, gym, multiagent, multiagent-reinforcement-learning, pettingzoo, ppo, reinforcement-learning, torch
- Language: Python
- Homepage:
- Size: 112 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE.md
Awesome Lists containing this project
README
# DeepRL
Implementation of some Deep Reinforcement Learning algorithms and environments.
## Description
The goal of this project is to have complete modularity with the algorithms and models used.
The implementations are completely made in PyTorch.
The environments used can either be single-agent using the Gymnasium
API, or multi-agents using the PettingZoo Parallel API. Most algorithms also support action masking.
## Getting Started
### Technologies used
* Python 3.11
* PyTorch 2.1.0
* Install all the requirements using `pip install -r requirements.txt`
### Usage
* Change the algorithm and the environment in the `main.py` file.
### Algorithms
The following algorithms are currently available:
* PPO (discrete [supports action masking] and continuous actions)
* A2C (discrete [supports action masking] and continuous actions)
* DQN (discrete actions)
### Environments
The following environments have been implemented:
* Snake
* Minesweeper
Any Gymnasium Env or PettingZoo ParallelEnv can be used.
## Authors
* Timothé Watteau (@timothewt)
## License
This project is licensed under the MIT License - see the LICENSE.md file for details