Getting started with Bonseyes and Docker
The Bonseyes Marketplace provides you with the tools and resources needed for supporting your workflow in machine learning in a collaborative manner, interacting with both direct colleagues as well as experts. The Marketplace supports you in both the practical aspects of sharing and collaborating on machine learning artefacts as well as discovering new datasets, advanced machine learning models and people to collaborate with.
This article describes how to easily get started with using the Bonseyes Marketplace and begin exchanging and collaborating of machine learning artefacts. In order to use the Bonseyes Marketplace you need to fulfilled the following pre-requisites:
- Have a reasonably new instance of Docker installed and running
- Configure Docker to support private repositories
- Add the private Bonseyes Marketplace Docker repository
The below instructions help you get up and running and submitting your first artefact!
Configuring Docker to support private repositories
For now, the Bonseyes Marketplace depends on unsecure repositories, while this is rectified in future revisions, it is currently necessary to configure the Bonseyes repository manually. To configure an unsecure repository, using the GUI, simply find the Setting tab, on Windows, or the Properties tab, under OSX. On Windows, go to Docker Daemon à Settings à Daemon, while on OSX the path is Docker Daemon à Preferences à Daemon. Here you may easily add the insecure repository:
bonseyes.bht.se:5000
Alternatively, the repository can be configured through the command line utility tool, by editing the Docker daemon configuration file by adding the following text:
{
"insecure-registries”: ["bonseyes.bth.se:5000"]
}
After restarting the Docker daemon, you can verify that everything is configure correctly by running:
$docker login Bonseyes.bht.se:5000
Using the Bonseyes Marketplace to exchange artefacts
In order to get started, an appropriate artefact needs to have been prepared, before it can be pushed to the repository. When the artefact is ready, tag the artefact, and push it to the repository like this:
docker tag my-artefact bonyeyes.bth.se:5000/my-artefact
and then execute the actual transmission of the artefact from your local computer to the repository using:
docker push Bonseyes.bth.se:5000/my-artefact
To download an artefact, after identifying a suitable artefact on the Bonseyes Marketplace, when receiving a direct link from a colleague or simply browsing the Bonseyes repository, the processing is highly similar. Instead of using the Docker pull command, simply replace it with the Docker push command, like this:
docker pull bonyeyes.bth.se:5000/my-artefact
Finally, you should be able to exchange your artefacts with whomever you wish, simply by providing them with a direct link to the repository and, if necessary, to this guide, and you are all set to go to collaborate, exchange and improve on your work.
Links
https://docs.docker.com/registry/
Keywords
bonseyes, data marketplace, machine learning, getting started