Completed
[] Pricing
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Vertex AI Workshop - Training, Serving, and Pipelines
Automatically move to the next video in the Classroom when playback concludes
- 1 [] Introduction to Sascha Heyer
- 2 [] Code, article, and videos
- 3 [] This episode's topics
- 4 [] Training ML Models
- 5 [] Training with Vertex AI
- 6 [] Training application
- 7 [] Why Container
- 8 [] Overall process
- 9 [] Demo
- 10 [] Vertex AI Training
- 11 [] Training Application
- 12 [] Enable monitoring for new versions of models
- 13 [] Using spot instances while kicking off the training jobs
- 14 [] Enabling Tensorflow real-time access while the job is traing
- 15 [] When to use Vertex AI vs when to use Google AI platform
- 16 [] Same components with Kubeflow
- 17 [] Control inside VPC
- 18 [] Serving ML Models
- 19 [] Different ways in Serving ML Models
- 20 [] Pre-build container for prediction
- 21 [] Custom container for prediction
- 22 [] Model serving steps
- 23 [] Upload model
- 24 [] Endpoint
- 25 [] Deploy model
- 26 [] Container requirements
- 27 [] Build customer container I
- 28 [] Build customer container II
- 29 [] Build customer container III
- 30 [] Getting predictions
- 31 [] Serving notebook demo
- 32 [] Optimizations around speeding up deployment
- 33 [] Working with Sagemaker relating to Vertex
- 34 [] Payload limitations
- 35 [] Limitations
- 36 [] Pricing
- 37 [] Machine Learning Teams don't need Kubernetes
- 38 [] Google Vertex AI Pipelines, a serverless product to run Kubeflow or TFX Pipelines
- 39 [] Vertex Pipelines and Kubeflow
- 40 [] Basic Pipeline
- 41 [] Required Modules
- 42 [] Components
- 43 [] Compiler
- 44 [] Demo
- 45 [] Component types
- 46 [] Predefined components
- 47 [] Component Specification
- 48 [] Share Components
- 49 [] Parameters
- 50 [] Model Lineage
- 51 [] Using Vertex Experiments
- 52 [] Scheduling pipelines
- 53 [] Production models trained and deploy
- 54 [] Vertex Batch prediction Service
- 55 [] Batch predictions are useful
- 56 [] Wrap up
