Skip to content

Installation on localhost

This guide shows how to install ML Aide on your local environment. This guide is only for demo purposes and should not be used for a production setup.

ML Aide is shipped with Docker. Therefore you need Docker installed. On Linux, you also have to install Docker Compose.

Clone Repo

Clone the ML Aide git repository

git clone https://github.com/MLAide/MLAide.git

For the installation, you don't have to compile or build anything. We only need the docker-compose.yaml from the demo/ directory.

Prepare

Add the following entry to your hosts file, to make sure keycloack can verify the user tokens. The file is located at /etc/hosts (Unix) or C:\Windows\System32\drivers\etc\hosts (Windows).

127.0.0.1 keycloak.mlaide

On Unix systems, you can use the following command to add the entry to your hosts file.

echo '127.0.0.1 keycloak.mlaide' | sudo tee -a /etc/hosts

Start

After that start ML Aide from your demo/ folder that is located in the cloned repository using:

cd demo
docker-compose up

Verify

Now you should have several containers running:

  • MLAide web UI
  • MLAide webserver
  • MongoDB which is used by the webserver to store all structured metadata
  • min.io that is used by the webserver to store all artifacts and models
  • keycloak to provide an identity provider, authentication, and authorization

Use

You can access the web UI on localhost:8880 with your browser. This demo provides three pre-defined users:

  • adam (password = adam1, email = adam@demo.mlaide.com)
  • bob (password = bob1, email = bob@demo.mlaide.com)
  • eve (password = eve1, email = eve@demo.mlaide.com)

Using the Python client

To track your machine learning experiments you can use the Python client. Start the Tutorial for your first steps.