=================================================================

Copyright and License

All Code in this repository - unless otherwise stated in local license or code headers is

Copyright 2023 Max Planck Institute for Intelligent Systems

Licensed under the terms of the GNU General Public Licence (GPL) v3 or higher. See: https://www.gnu.org/licenses/gpl-3.0.en.html

installation

• create workspace

mkdir MultitaskRL
cd MultitaskRL

• setup isaac-gym

1. download isaac-gym from https://developer.nvidia.com/isaac-gym
2. extract isaac-gym to the workspace
3. create conda environment and install the dependencies

bash IsaacGym_Preview_4_Package/isaacgym/create_conda_env_rlgpu.sh 
conda activate rlgpu

• clone this repository to the workspace and install dependencies

git clone https://github.com/robot-perception-group/concurrent_composition.git
pip install -r concurrent_composition/requirements.txt

Run experiments

• enter the RL workspace

cd concurrent_composition/

• The experiments are stored in the sweep folder. For example, hyperparameter tuning for dac in Pointmass environment

wandb sweep sweep/dac_ptr_transfer.yml --count 25

The experimental results will gather in Wandb.

• To train a agent

python3 run.py agent=DACGPI env=Pointer2D env.save_model=True

• To play a trained agent, first specify the path to the model in play_ptr2d.py, then run the command

python3 play_ptr2d.py agent=DACGPI env=PointMass2D env.sim.headless=False env.num_envs=1