Table of Contents

1. About The Project
2. Usage
3. TODO
4. License
5. Contact
6. Acknowledgements

About The Project

Robot learning projects are usually tested in a system with a GUI interface. After the test, we want to deploy the project on a server for large-scale training to improve training efficiency. However, the local training environment may cause hardware compatibility issues during migration and deployment. I want to develop a Pybullet robot learning container based on Docker.

Here's why:

• Portability: we can package pybullet project and dependencies into a portable image and then publish to any popular Linux or Windows machine.
• Efficient: Through Docker, you can control the server's CPU or GPU resources and run multiple containers for training simultaneously.

A list of commonly used resources that I find helpful are listed in the acknowledgements.

Built With

This section should list any major frameworks that you built your project using. Leave any add-ons/plugins for the acknowledgements section. Here are a few examples.

• Docker
• Pybullet
• anaconda3 dockerfile
• wandb

Usage

This is an example of how you may give instructions on setting up your project locally. To get a local copy up and running follow these simple example steps.

Prerequisites

Install docker at https://docs.docker.com/get-docker/

Docker pull (recommend)

Run with CPU

docker pull zanehuang/pybullet_spinningup:cpu
docker run -i -v "$(pwd)"/src:/src -t zanehuang/pybullet_spinningup:cpu /bin/bash

Build image from dockerfile

1. Clone the repo

git clone https://github.com/zebin-huang/docker_pybullet.git

2. Build docker image from dockerfile. It will take some minutes to build the image.

docker build -t your_name/pybullet_spinningup:cpu .

3. Run the docker image

mkdir src
docker run -i -t your_name/pybullet_spinningup:cpu -v "$(pwd)"/src:/src /bin/bash

4. Pip install spiningup, you can replace your env

bash
conda activate spinningup
# clone my version, I made some changes.
git clone https://github.com/borninfreedom/spinningup.git
cd spinningup
pip install -e .

TODO

• wandb support
• GPU image development
• Docker multiple container deployment
• Docker CPU/GPU resource management
• Remove unnecessary dependencies

License

Distributed under the Apache License 2.0 License. See LICENSE for more information.

Contact

Your Name - @Zebin Huang

Project Link: https://github.com/zebin-huang/docker_pybullet

Acknowledgements

• borninfreedom/kuka-reach-drl

• othneildrew/Best-README-Template