MLOps for R with Azure Machine Learning
summary
Demonstrates end-to-end MLOps for R using Azure Machine Learning service and the azuremlsdk for R to train, version and deploy models. Uses an RStudio compute instance (DS2_v2 two‑core, ~7GB RAM) and optional GPU node (six cores, ~50GB RAM), submits caret jobs (glm, knn, glmnet) to a cluster scaled min nodes 1 max nodes 2, tracks experiments, registers the best model, deploys to Azure Container Instances and AKS, exposes a REST endpoint consumed by a Shiny app, and integrates CI/CD with Azure Pipelines/GitHub Actions.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 1:18What is DevOps
- 2:44MLOps vs DevOps
- 7:10Azure Machine Learning Service
- 10:05Azure ML SDK
- 11:17Example
- 12:10Create a compute instance
- 13:46Create a ML compute cluster
- 14:34Training models
- 15:50Results
- 16:11Versioning
- 16:24Deploying
- 16:48Demo
- 17:32Rest endpoint
- 18:36Summary
- 19:27Wrapup
related talks
48mStreamlining your Machine Learning projects with Azure Machine Learning
44mPractical deep dive into machine learning techniques and MLOps
42mPython SDK to train, package and deploy models
10mGet Started with Azure Machine Learning
43mA lap around Azure Machine Learning services
27mBeginners guide to Azure MLOps
47mPractical deep dive into machine learning techniques and MLOps Q&A
45mMLOps and ML-Flops: going from dev to prod