PaCo: Parameter-Compositional Multi-Task Reinforcement Learning (NeurIPS 2022)
Introduction
This repo is for the Parameter-Compositional MTRL method proposed in "PaCo: Parameter-Compositional Multi-Task Reinforcement Learning".
Usage
Installation
The training environment (PyTorch and dependencies) can be installed as follows:
git clone --recursive https://github.com/TToTMooN/paco-mtrl.git
cd paco-mtrl
pip install -e . -e alf
Training
cd paco/examples/
CUDA_VISIBLE_DEVICE=0 python3 -m alf.bin.train --conf=metaworld_mt10_paco_conf.py --root_dir=EXP_PATH
More examples will be added soon.
Contact
For questions related to PaCo, please send me an email: lingfengsun@berkeley.edu