mlflow

A self-hosted experiment tracking service

Getting Started

These instructions will give you a copy of the project up and running on your local machine for development and testing purposes.

Prerequisites

Requirements for the software and other tools to build, test and push

• Git
• Kubernetes command line tool (kubectl)
• Helm
• Configured kubernetes cluster

Supported IDEs

You can use the IDE of your choice to continue developing the project. However, the following IDEs are recommended and officially supported:

• VSCodeium
• Visual Studio Code

Secret Management

Sensitive data is stored in argo cd.

Deployment

As soon as changes are pushed to the main branch on the remote repository, the new version will be deployed.

You have to build and push the docker image manually with the follwoing code:

docker login -u <username> --password-stdin git.testgoofy.ch
docker build -t git.testgoofy.ch/homeserver/mlflow:<version> -t git.testgoofy.ch/homeserver/mlflow:latest .
docker push git.testgoofy.ch/homeserver/mlflow:<version>
docker push git.testgoofy.ch/homeserver/mlflow:latest

Development

This repository is a full functional helm chart. Therefore you can use the helm command line tool to deploy a development instance of the project. To deploy a development instance, run the following command:

helm upgrade --install --namespace staging --create-namespace mlflow . \
--set host=mlflow.dev.testgoofy.ch \
--set imageCredentials.username=<Redacted>
--set imageCredentials.password=<Redacted>
--set imageCredentials.email=<Redacted>
--set postgres.password=<Redacted>

Authors

• Lukas Gysin - Main Contributor & Project Owner

License

This project is licensed under the MIT License - see the LICENSE.md file for details

Acknowledgments

• Billie Thompson - Provided README Template
• Contributor Covenant - Provided CODE_OF_CONDUCT Template