.md .pdf

repository open issue suggest edit

Getting-Started Tutorial

This tutorial provides a hands-on introduction to using MLRun to implement a data science workflow and automate machine-learning operations (MLOps).

The tutorial covers MLRun fundamentals such as creation of projects and data ingestion and preparation, and demonstrates how to create an end-to-end machine-learning (ML) pipeline. MLRun is integrated as a default (pre-deployed) shared service in the Iguazio MLOps Platform.

You can see another example which demonstrate how to Convert Research Notebook (NYC Taxi) to Operational Pipeline. For additional demos refer to MLRun demos repository at github.com/mlrun/demos.

In this tutorial you’ll learn how to:

• Collect (ingest), prepare, and analyze data

• Train, deploy, and monitor an ML model

• Create and run an automated ML pipeline

You’ll also learn about the basic concepts, components, and APIs that allow you to perform these tasks, including

• Setting up MLRun

• Creating and working with projects

• Creating, deploying and running MLRun functions

• Using MLRun to run functions, jobs, and full workflows

• Deploying a model to a serving layer using serverless functions

The tutorial is divided into four parts, each with a dedicated Jupyter notebook. The notebooks are designed to be run sequentially, as each notebook relies on the execution of the previous notebook:

• Part 1: MLRun Basics
  • Introduction to MLRun
  • Step 1: Installation and Setup
  • Step 2: Creating a Basic Function
  • Step 3: Running the Function on a Cluster
  • Step 4: Viewing Jobs on the Dashboard (UI)
  • Step 5: Scheduling Jobs
  • Done!
• Part 2: Training an ML Model
  • Prerequisites
  • Step 1: Setup and Configuration
  • Step 2: Creating a Training Function
  • Step 3: Exploring the Data with an MLRun Marketplace Function
  • Step 4: Testing Your Model
  • Done!
• Part 3: Serving an ML Model
  • Prerequisites
  • Step 1: Setup and Configuration
  • Step 2: Writing A Simple Serving Class
  • Step 3: Deploying the Model-Serving Function (Service)
  • Step 4: Using the Live Model-Serving Function
  • Step 5: Viewing the Nuclio Serving Function on the Dashboard
  • Done!
• Part 4: Projects and Automated ML Pipeline
  • Prerequisites
  • Step 1: Setting Up Your Project
  • Step 2: Updating Project and Function Definitions
  • Step 3: Defining and Saving a Pipeline Workflow
  • Step 4: Registering the Workflow
  • Step 5: Running A Pipeline
  • Step 6: Viewing the Pipeline on the Dashboard (UI)
  • Step 7: Invoking the Model
  • Done!

Quick-Start Guide Part 1: MLRun Basics