Amir Ardalan Kalantari Dehagh - 260851609 Oleg Zhilin - 260581713

COMP 767 Final Project - Policy Adjustment in Robotics

Note that when cloning this repository, the saved models are included so it might take a bit of time (the total size is slightly over 1GB)

Instructions

• Create a virtual environment: python -m venv venv
• Activate it: source venv/bin/activate
• Install dependencies: pip install -r requirements.txt
• Add custom environments as dependencies: cd analysis/continuous_cartpole; pip install -e .
• Finally, run one of the following files from the analysis directory:
  • cartpole.py: to train PILCO models (takes a long time)
  • cartpole_from_file.py: visualize learning of PILCO models in the saved directory
  • train_transfer.py: train a set of transfer models for each PILCO iteration in the saved directory
  • path_analysis.py: to reproduce path plots from the report (needs saved and transfer-save directory contents)