Residual Reinforcement Learning used for insertion
tensorflow==1.14.0
Pillow
stable_baselines
It is recommended to use a virtual environment.
Install requirements with
pip install -r requirements.txt
Install gym wrapper
cd gym-insertion
pip install -e .
Install the simulation, either for Godot (https://github.com/hoel-bagard/InsertionGodot) or Unity (https://github.com/xc-jp/insertion-simulation)
After installing the insertion simulation and starting it, run
python train.py