Machine Learning Model Deployment Guide

This repository contains my guide to deploy ML models. The repository was forked from

deploying-machine-learning-models

which contains the companion code of the Udemy course Deployment of Machine Learning Models by Soledad Galli & Christopher Samiullah.

The guide or notes for my future self done after following the course are in: ./ML_Deployment_Guide.md.

Notes on the Contents of the Repository

• Presentations: provided as a Dropbox download link, located in ./udemy_ml_deployment/deployment_of_ML_presentations; but not committed.
• Datasets: downloaded from kaggle: House Prices - Advanced Regression Techniques; located in ./data/house-prices-advanced-regression-techniques; but not committed.

Overview of the Contents of the Course

1. Overview of Model Deployment
2. Machine Learning System Architecture
3. Research Environment: Developing a Machine Learning Model
4. Packaging the Model for Production
5. Serving and Deploying the Model via REST API - FastAPI
6. Continuous Integration and Deployment Pipelines - CicleCI
7. Deploying the ML API with Containers
8. Differential Tests
9. Deploying to IaaS (AWS ECS)

How to Start?

Continue in ./ML_Deployment_Guide.md for the detailed guide/notes.

Also check these links:

• Through deployment example: census_model_deployment_fastapi.
• My personal notes on the Udacity MLOps nanodegree: mlops_udacity. The module guide about deployment: MLOpsND_Deployment.md

Relevant Links

• Check this example deployment: census_model_deployment_fastapi
• Check my personal notes on the Udacity MLOps nanodegree: mlops_udacity:
  • The module guide about deployment: MLOpsND_Deployment.md
  • Example and exercise repository related to the topic of deployment: mlops-udacity-deployment-demos.
• My guide on CI/DC: cicd_guide
• My boilerplate for reproducible ML pipelines using MLflow and Weights & Biases: music_genre_classification.
• A very simple Heroku deployment with the Iris dataset and using Flask as API engine.
• Notes on how to transform research code into production-level packages: customer_churn_production.
• My summary of data processing and modeling techniques: eda_fe_summary.

Authorship

Notes by Mikel Sagardia, 2022.
No guarantees.