Serverless Machine Learning 101
summary
Shows how to serve a retrained InceptionV3 image classifier (cats vs dogs) as an HTTP POST endpoint using Azure Functions and TensorFlow with a Python requirements file and VS Code + Functions extension. Claims pay‑per‑use scaling, 10–30 minute runtime limits (use durable functions for longer jobs), and stateless short‑lived execution. Demo uses transfer learning, installs tensorflow/pillow/numpy, attempts local debug (fails) but successfully packages and deploys to West Europe — model hosting works after deployment.
AI-written summary · about
chapters (9)
- 0:00Start
- 3:37Why Serverless computing?
- 5:42Microsoft Azure functions
- 6:49You - the developer
- 9:52Serverless characteristics
- 13:32Good use cases for Faas
- 15:21Some ML systems characteristics
- 17:04The Scenario
- 18:21Requirements
related talks
45mServerless Machine Learning: a beginner's guide
39mServerless at the end of the Universe
39mMachine Learning Made Easy on Kubernetes. DevOps for Data Scientists
41mBuilding Serverless Machine Learning Models in the Cloud
15mServerless Tensorflow 1.0.0 on AWS. Wait what?
31mBuild an AI-powered Pet Detector in Visual Studio Code
18mServerless AI's Rise: Revolutionizing ML Deployment
12mServerless Deep Learning