README

If you prefer, you can watch this video tutorial.

1. Requirements
2. Installing Docker in a Debian-based Linux distribution
3. Setting up the toolkit
4. Running the toolkit
5. How can I create and send my demos to this project?

What is this repository for?

This repository contains scripts to set up and run the Docker Containers which launch BahiaRT's Setplays Learning from Demonstration Toolkit. The toolkit is a set of a fork of RoboViz 1.6.0 modified to support the Demonstration Mode and a Fork of Strategy Planner (SPlanner) to support Demonstration Mode and new behaviors and features.

See USAGE.md to learn how to use this tool kit and send your contribution to this research project.

Requirements

• Install git and docker in a modern Linux distribution.
• Use of Docker Desktop in Windows 10 or + is also possible, but we haven't tested this setup. If you want to try using the toolkit in Windows, install and set up Docker Desktop, WSL2, and an X Server for Windows.
• Check if your computer uses an NVIDIA Graphic Card and install updated NVIDIA Drivers. Make sure your OS is really using your NVIDIA drivers.
• Just update your graphic card drivers if you do not use an NVIDIA Graphic Card.

Installing Docker in a Debian-based Linux distribution

$ sudo apt-get update

$ sudo apt-get install -y docker.io

$ sudo gpasswd -a {user-name} docker

Replace {user-name} by your user name in the Linux system.

For other distributions, replace apt-get with your command-line package manager(e.g., yum, zypper).

Setting up the toolkit

• First, clone this repository:

git clone https://bitbucket.org/bahiart3d/setplaysdataset.git

• Change your current folder to the repository:

cd setplaysdataset

• If you are using a Debian-based linux distribution, just run setup.sh:

./setup.sh nvidia

or

./setup.sh mesa

If you use NVIDIA choose the first option. Otherwise use mesa as the argument. If you do not use a Debian-based Linux distribution, edit setup.sh and replace apt-get with your command-line package manager (e.g., yum, zypper, etc.)

• If you are trying in a Windows 10 or + without an NVIDIA Graphics card, run setup.cmd in the Windows Powershell:

setup.cmd

We could not correctly set up NVidia in the Docker Desktop setup. But you can try it.

The setup can take several minutes depending on your internet bandwidth. But, after setup, it is very fast to run the toolkit.

Running the toolkit

• For systems using NVIDIA drivers:

./run.sh nvidia

• For other systems:

./run.sh mesa

• For Windows systems without NVIDIA Graphics Cards, run in the Powershell:

run.cmd