Giter Site home page Giter Site logo

chloe4d / marl_transfer Goto Github PK

View Code? Open in Web Editor NEW

This project forked from sumitsk/marl_transfer

0.0 1.0 0.0 10.58 MB

Code for paper 'Learning transferable cooperative behaviors in multi-agent teams' (ICML 2019)

Home Page: https://graphreason.github.io/papers/29.pdf

License: MIT License

Python 100.00%

marl_transfer's Introduction

Learning Transferable Cooperative Behavior in Multi-Agent Teams

Akshat Agarwal*, Sumit Kumar*, Katia Sycara

Robotics Institute, Carnegie Mellon University

This is the official repository of the 'Learning Transferable Cooperative Behavior in Multi-Agent Teams' paper, available at https://arxiv.org/abs/1906.01202

This work has been presented at the Learning and Reasoning with Graph-Structured Reprsentations workshop (https://graphreason.github.io/papers/29.pdf) at ICML, 2019 held in Long Beach, USA.

Installation

See requirements.txt file for the list of dependencies. Create a virtualenv with python 3.5 and setup everything by executing pip install -r requirements.txt.

Examples

See arguments.py file for the list of various command line arguments one can set while running scripts.

Normal Training

Training on Coverage Control (simple_spread) environment can be started by running:

python main.py --env-name simple_spread --num-agents 3 --entity-mp --save-dir 0

Similarly scripts for Formation Control (simple_formation) and Line Control (simple_line) can be launched as:

python main.py --env-name simple_formation --num-agents 3 --save-dir 0

python main.py --env-name simple_line --num-agents 3 --save-dir 0

Specify the flag --test if you do not want to save anything.

Curriculum Training

To start curriculum training, specify the number of agents in automate.py file and execute:

python automate.py --env-name simple_spread --entity-mp --save-dir 0

Results

The models trained via curriculum learning on the three environments can be found in models subdirectory.

The corresponding results obtained from the trained policies are located in videos subdirectory.

Transfer

You can also continue training from a saved model. For example, for training a team of 5 agents in simple_spread task from a policy trained with 3 agents, execute:

python main.py --env-name simple_spread --entity-mp --continue-training --load-dir models/ss/na3_uc.pt --num-agents 5

Contact

For any queries, feel free to raise an issue or contact the authors at [email protected] or [email protected].

License

This project is licensed under the MIT License.

marl_transfer's People

Contributors

agakshat avatar sumitsk avatar

Watchers

 avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.