Installation And Setup

1. sudo apt install xvfb (for training headless)
2. git clone --recurse-submodules https://github.com/xli4217/lof_robot_learning.git
3. Download CoppeliaSim_Edu_V4_1_0_Ubuntu18_04/ at https://www.coppeliarobotics.com/downloads# in the lof_robot_learning directory
4. Follow https://github.com/stepjam/PyRep to install PyRep (installed in libs/PyRep)
5. conda env create -f env.yml
6. source setup_path.sh

Train And Run

1. conda activate lof
2. To start training run python src/train.py (by default training is done headless, if visualization of the training is desired, in src/train.py modify HEADLESS=False )
3. To test trained policy run python src/rollout.py
4. To plot learning curve, in package root directory run python -m spinup.run plot experiments/<experiment name>/