MLRun has many code examples and tutorial Jupyter notebooks with embedded documentation, ranging from examples of basic tasks to full end-to-end use-case applications, including the following; note that some of the examples are found in other mlrun GitHub repositories:
- Learn MLRun basics — mlrun_basics.ipynb
- Convert local runs to Kubernetes jobs and create automated pipelines in a single notebook — mlrun_jobs.ipynb
- End-to-end ML pipeline— demos/scikit-learn, including:
- Data ingestion and analysis
- Model training
- Verification
- Model deployment
- MLRun with scale-out runtimes —
- Distributed TensorFlow with Horovod and MPIJob, including data collection and labeling, model training and serving, and implementation of an automated workflow — demos/image-classification-with-distributed-training
- Serverless model serving with Nuclio — xgb_serving.ipynb
- Dask — mlrun_dask.ipynb
- Spark — mlrun_sparkk8s.ipynb
- MLRun project and Git life cycle —
- Load a project from a remote Git location and run pipelines — load-project.ipynb
- Create a new project, functions, and pipelines, and upload to Git — new-project.ipynb
- Import and export functions using files or Git — mlrun_export_import.ipynb
- Query the MLRun DB — mlrun_db.ipynb
- Additional end-to-end use-case applications — mlrun/demos repo
- MLRun functions Library — mlrun/functions repo