Skip to content
This repository has been archived by the owner on Dec 16, 2022. It is now read-only.

Latest commit

 

History

History
38 lines (33 loc) · 3.08 KB

README.md

File metadata and controls

38 lines (33 loc) · 3.08 KB
Raw

A Comparative Evaluation of Deep Reinforcement Learning Frameworks

This repository contains the implementation part of our work in which we compared several deep reinforcement learning frameworks: Dopamine, Horizon and Ray.

How to use

Using Docker (recommended)

  1. Install Docker following the instructions from here.
  2. Install nvidia-docker from here to run experiments with GPU inside the container.
  3. Build necessary Docker images by executing the following script inside the repository folder:
bash ./scripts/docker/init.sh
  1. Once the images are built, start Postgres and project containers by running bash ./scripts/docker/start.sh. When the script is done, you'll have access to the interactive terminal inside the project container.
  2. Once inside the container, you can validate the setup by running bash ./scripts/test.sh.
  3. Alternatively, proceed with the experiments by executing bash ./scripts/evaluation/{environment}/eval_all.sh for the complete evaluation or bash ./scripts/evaluation/{environment}/eval_{framework}.sh inside the container for evaluating single frameworks.
  4. Run tensorboard --logdir=results inside the container, and view the Tensorboard summary of evaluation results in your local browser at localhost:6006.

Install from source (development)

  1. Install Anaconda from here (make sure to download the Python 3 version). Leave Anaconda's installation directory default (home/miniconda3).
  2. Check Anaconda's version by executing conda -V. If Anaconda's version is <=4.6.8, run conda update conda.
  3. Create project environment by running bash ./scripts/setup_local_env.sh.
  4. You can validate your setup by running bash ./scripts/test.sh.
  5. If you'd like to evaluate the frameworks against the Park's Query Optimizer environment, then do the following extra steps:
    • edit /etc/hosts file and add the following alias: 127.0.0.1 docker-pg;
    • edit /etc/hosts file and add the following alias: 127.0.0.1 drl-fw;
    • build the Postgres Docker image used by the environment: docker build -t pg park/query-optimizer/docker/;
    • start the Postgres container: docker start docker-pg || docker run --name docker-pg -p 0.0.0.0:5432:5432 --net drl-net --privileged -d pg.
  6. Run experiments for a specific environment by executing bash ./scripts/evaluation/{environment}/eval_all.sh. Alternatively, you can run experiments for individual frameworks by running bash ./scripts/evaluation/{environment}/eval_{framework}.sh.
  7. View evaluation results in Tensorboard (localhost:6006) after running tensorboard --logdir=results. You may need to activate the proper environment first (conda activate drl-frameworks-env).