wensun / Imitation-Learning-from-Observation

Forward Adversarial Imitation Learning

This repository implements the algorithms presented in the paper (published in ICML 2019)

Dependencies

• We advise the reader to use virtualenv so that installing dependencies is easy
• Install the gym package given with the code. To do this, activate your virtual env and go inside gym/ and run pip install -e .
• Install tensorflow (version 1.12.0)
• Install the OpenAI baselines package given with the code. To do this, activate your virtual env and go inside baselines/ and run pip install -e .
• Install other dependencies. To do this, activate your virtual env and go inside FAIL/ and run pip install -r requirements.txt
• Install Mujoco-py: pip install -U 'mujoco-py<1.50.2,>=1.50.1'
• We expect the reader to have a mujoco license to run the mujoco experiments in the paper

Running the code

To download datasets, go to the directory FAIL/

• run python download_datasets.py

To run the experiments, go to the directory FAIL/ and for the environment

• Fetchreach, run ./scripts/fetchreachexperiments.sh
• Swimmer, run ./scripts/swimmerdiscreteexperiments.sh
• Reacher, run ./scripts/reacherdiscreteexperiments.sh

The results are generated and stored in the FAIL/data folder (for all the 10 random seeds in each experiment for every method)