The README for the original mjrl package, as used in this work, is copied below.
This package contains implementations of various RL algorithms for continuous control tasks simulated with MuJoCo.
The main package dependencies are python=3.5
, gym
, mujoco-py
, and pytorch
. See setup/README.md
(link) for detailed install instructions.
If you find the package useful, please cite the following papers.
@INPROCEEDINGS{Rajeswaran-NIPS-17,
AUTHOR = {Aravind Rajeswaran and Kendall Lowrey and Emanuel Todorov and Sham Kakade},
TITLE = "{Towards Generalization and Simplicity in Continuous Control}",
BOOKTITLE = {NIPS},
YEAR = {2017},
}
@INPROCEEDINGS{Rajeswaran-RSS-18,
AUTHOR = {Aravind Rajeswaran AND Vikash Kumar AND Abhishek Gupta AND
Giulia Vezzani AND John Schulman AND Emanuel Todorov AND Sergey Levine},
TITLE = "{Learning Complex Dexterous Manipulation with Deep Reinforcement Learning and Demonstrations}",
BOOKTITLE = {Proceedings of Robotics: Science and Systems (RSS)},
YEAR = {2018},
}
This package is maintained by Aravind Rajeswaran and other members of the Movement Control Lab, University of Washington Seattle.