Skip to content

Pytorch based library containing reinforcement learning agents with forward models and intrinsic motivation modules

License

Notifications You must be signed in to change notification settings

Iskandor/MotivationModels

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Seer for reinforcement learning agents

Repository contains reinforcement learning agents (A2C,DQN,DDPG,PPO), different forward models and artificial motivation modules from the knowledge-based category of intrinsic motivation. Models are being tested on robotic environments and Atari games.

Installation

Prerequisites are c++ compiler and swig then proceed with installation:

$ git clone https://github.com/Iskandor/MotivationModels.git
$ cd MotivationModels
$ pip install -r requirements.txt

Usage

$ python main.py --algorithm alg_name --env env_name --config config_id

Avaialable algorithms are ppo, ddpg, a2c, dqn

Environment names and ids of configuration can be found in config/alg_name.config.json

eg. ddpg.config.json

Development

Project is still in development and we are planning to implement additional models and methods of reinforcement learning and intrinsic motivation.

Author

Matej Pechac is doctoral student of informatics specializing in the area of reinforcement learning and intrinsic motivation

About

Pytorch based library containing reinforcement learning agents with forward models and intrinsic motivation modules

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages