Repo showing two approaches (one in two flavours) to structuring ML projects/pipelines using MLflow Projects. Additionally, MLflow Tracking used for tracking params, metrics and artifacts
- Install requirements
pip install -r requirements.txt
Move to repo root directory
cd <REPO_ROOT>
Execute to run simple project:
mlflow run simple --experiment-name simple
Execute to run multi-step project:
mlflow run multistep --experiment-name multistep
Execute to run multi-env project:
(requires docker - cd multienv/steps/1_docker & docker build -t mlflow_docker . and R)
mlflow run multienv --experiment-name multienv
Execute to start local MLflow server to see tracking UI:
mlflow server