Metadata-Version: 2.1
Name: marlben
Version: 1.0.2
Summary: Marlben is a multi-agent reinforcement learning benchmark based on the NeuralMMO game engine.
Author-email: JBR AI labs <ego.vladi@gmail.com>
License: MIT License
Project-URL: Homepage, https://github.com/jbr-ai-labs/marlben
Project-URL: Repository, https://github.com/jbr-ai-labs/marlben
Project-URL: Documentation, https://github.com/jbr-ai-labs/marlben/wiki
Project-URL: Bug Report, https://github.com/jbr-ai-labs/marlben/issues
Keywords: Multi-Agent Reinforcement Learning,game,RL,AI,Public Good
Classifier: Development Status :: 4 - Beta
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Intended Audience :: Science/Research
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.9
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: numpy (>=1.24.3)
Requires-Dist: pettingzoo (>=1.14.0)
Requires-Dist: tqdm (~=4.61.1)
Requires-Dist: scipy (~=1.8.0)
Requires-Dist: imageio (~=2.8.0)
Requires-Dist: Twisted (~=19.2.0)
Requires-Dist: autobahn (~=19.3.3)
Requires-Dist: vec-noise (~=1.1.4)
Requires-Dist: matplotlib (>=3.0)
Requires-Dist: openskill (~=4.0.0)
Requires-Dist: gymnasium (>=0.26.3)
Provides-Extra: all
Requires-Dist: ray[rllib] (~=2.5.0) ; extra == 'all'
Requires-Dist: torch (==1.13.1) ; extra == 'all'
Requires-Dist: wandb (~=0.15.4) ; extra == 'all'
Provides-Extra: rllib
Requires-Dist: ray[rllib] (~=2.5.0) ; extra == 'rllib'
Requires-Dist: torch (==1.13.1) ; extra == 'rllib'
Requires-Dist: wandb (~=0.15.4) ; extra == 'rllib'
Provides-Extra: testing
Requires-Dist: pytest (==7.1.1) ; extra == 'testing'

![logo](https://github.com/jbr-ai-labs/marlben/assets/22059171/cab386fb-3b49-4f02-b59a-49dd2de3cc9c)

[![PyPI version](https://badge.fury.io/py/marlben.svg)](https://pypi.org/project/marlben/1.0.1/)
![marlben-tests](../../actions/workflows/python-package-conda.yml/badge.svg)

Marlben is a multi-agent reinforcement learning benchmark based on the [NeuralMMO](https://github.com/NeuralMMO/environment) game engine

## Installation
https://github.com/jbr-ai-labs/marlben/wiki/Quick-Start#installation

### Installation via Pip
`Python3.9` is required

```
pip install marlben==1.0.1
```

## Usage

https://github.com/jbr-ai-labs/marlben/wiki/Quick-Start#launching-an-environment

### Examples
Examples are located in `examples` folder
- `rllib_integration.py` demonstrates how to start learning process with default config for `Corridor` env
- `override_config.py` demonstrates how to customize configs by inheriting from default class
- `custom_map_generator.py` demonstrates how to customize map generation for any environment
- `custom_environment.py` demonstrates how to customize whole environment structure by overriding all building blocks of the benchmark

## Wiki
https://github.com/jbr-ai-labs/marlben/wiki
