terminus7/mlflow repository overview

⁠MLFlow Tracking Server

This repo hold the creation of Docker image to run MLFlow Tracking Server.

This component is part of a toolkit used to simplify the data scientists daily work. For more details check out the Science Toolkit documentation⁠

⁠Run it

With the following command you can access to the MLFlow Tracking server opening a browser the URL http://localhost:5000⁠. The artifacts are stored in a Minio bucket and the logs in the local directory /mlflow/tracking, if you want to persist the log folder you need to mount an external volume to this path. The published Docker image is public in Docker Hub⁠.

docker run -p 5000:5000 \
        -e "MLFLOW_S3_ENDPOINT_URL=http://my_minio.url" 
        -e "AWS_ACCESS_KEY_ID=user" 
        -e "AWS_SECRET_ACCESS_KEY=pass" 
        -e "ARTIFACTS_BUCKET=mlflow-artifacts" 
        terminus7/mlflow:latest