https://github.com/jbr-ai-labs/mamba

This code accompanies the paper "Scalable Multi-Agent Model-Based Reinforcement Learning".
https://github.com/jbr-ai-labs/mamba

Last synced: 3 months ago
JSON representation

This code accompanies the paper "Scalable Multi-Agent Model-Based Reinforcement Learning".

• Host: GitHub
• URL: https://github.com/jbr-ai-labs/mamba
• Owner: jbr-ai-labs
• License: mit
• Created: 2022-01-20T09:19:56.000Z (about 3 years ago)
• Default Branch: main
• Last Pushed: 2023-07-05T18:01:19.000Z (over 1 year ago)
• Last Synced: 2024-08-02T13:23:15.840Z (7 months ago)
• Language: Python
• Size: 1.6 MB
• Stars: 43
• Watchers: 3
• Forks: 9
• Open Issues: 3
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# MAMBA

This code accompanies the paper "[Scalable Multi-Agent Model-Based Reinforcement Learning](https://arxiv.org/abs/2205.15023)".

The repository contains MAMBA implementation as well as fine-tuned hyperparameters in ```configs/dreamer/optimal``` folder.

## Installation

`python3.7` is required

```

pip install wheel

pip install flatland-2.2.2/

pip install -r requirements.txt 

```

Installing Starcraft:

https://github.com/oxwhirl/smac#installing-starcraft-ii

## Usage

```

python3 train.py --n_workers 2 --env flatland --env_type 5_agents

```

Two environments are supported for env flag: flatland and starcraft.

### Optimal parameters

To train agents with optimal parameters from the paper they should be copied from `configs/dreamer/optimal/` folder to [DreamerAgentConfig.py](https://github.com/jbr-ai-labs/mamba/blob/main/configs/dreamer/DreamerAgentConfig.py) and [DreamerLearnerConfig.py](https://github.com/jbr-ai-labs/mamba/blob/main/configs/dreamer/DreamerLearnerConfig.py)

## SMAC



The code for the environment can be found at 

[https://github.com/oxwhirl/smac](https://github.com/oxwhirl/smac)

## Flatland



The original code for the environment can be found at 

[https://github.com/jbr-ai-labs/NeurIPS2020-Flatland-Competition-Solution](https://github.com/jbr-ai-labs/NeurIPS2020-Flatland-Competition-Solution)

## Code Structure

- ```agent``` contains implementation of MAMBA 

  - ```controllers``` contains logic for inference

  - ```learners``` contains logic for learning the agent

  - ```memory``` contains buffer implementation

  - ```models``` contains architecture of MAMBA

  - ```optim``` contains logic for optimizing loss functions

  - ```runners``` contains logic for running multiple workers

  - ```utils``` contains helper functions

  - ```workers``` contains logic for interacting with environment

- ```env``` contains environment logic

- ```networks``` contains neural network architectures

## Citation

    @inproceedings{10.5555/3535850.3535894,

          author = {Egorov, Vladimir and Shpilman, Alexei},

          title = {Scalable Multi-Agent Model-Based Reinforcement Learning},

          year = {2022},

          isbn = {9781450392136},

          publisher = {International Foundation for Autonomous Agents and Multiagent Systems},

          address = {Richland, SC},

          booktitle = {Proceedings of the 21st International Conference on Autonomous Agents and Multiagent Systems},

          pages = {381–390},

          numpages = {10},

          keywords = {communication, multi-agent reinforcement learning, model-based reinforcement learning},

          location = {Virtual Event, New Zealand},

          series = {AAMAS '22}

    }