Azure ML, Azure Kubernetes Service, Model deployment

on YouTube ↗24m

" summary

Builds a decision tree regressor on a wine dataset inside an Azure ML workspace that auto-creates storage and App Insights, runs three experiment runs on a d2 v2 compute cluster, registers the model, and deploys to an AKS inference cluster with a real time endpoint that returns HTTP 200 and predicts quality 7 for a sample. Uses an inference script, a conda environment, and Application Insights logs to monitor requests and predictions.

AI-written summary · about

" chapters (12)

1. 0:00Introduction
2. 2:45Azure ML
3. 4:55Model building
4. 7:05Compute cluster
5. 8:30Create training script
6. 10:30Create estimator
7. 11:20Submit experiment
8. 12:05Submit runs
9. 13:40Model registration
10. 16:00Dependencies
11. 18:00Deployment
12. 21:35Logs

" related talks

• 48mMachine Learning in Azure DatabricksTerry McCann · SQLBits 2020
• 57mTaking Models to the Next Level with Azure Machine LearningHenk Boelman · NDC Sydney 2020
• 58mDeep Dive into Azure Automated Machine LearningVivek Raja P S · Global Summer 2020 Azure Data Fest 2020
• 01h 05mAzure Machine Learning for the DeveloperPeter Myers · NDC Sydney 2016
• 49mBuilding Advance Analytics pipelines with Azure DatabricksLace Lofranco · NDC Sydney 2018
• 59mPredicting the future as a service with Azure ML and RBarbara Fusinska · NDC London 2017
• 49mAutomated Machine Learning on AzureGavita Regunath · SQLBits 2023
• 43mA lap around Azure Machine Learning servicesPeter Drougge · Update Conference Prague 2018