Files

Kubeflow pipeline on Vertex

Overview

Theses examples show the codes of KFP v2 with aiplatform SDK

  • create endpoint
  • train model
  • deploy model

Requirements

  • Python 3.8
  • Vertex on Google Cloud Platform

Quickstart

  • Compiler
    • use python to compiler example.py to example.json
  • Run
    • Open Vertex on GCP
    • Click 'Pipelines' and create run
    • Click 'Upload file' to upload example.json
    • Select bucket then submit

Cleanup

Following are optional to cleanup

  • endpoint
    • Click 'Online prediction'
    • Click 'Actions' to delete endpoint
  • model
    • Click 'Model Registry'
    • Click 'More' to delete model