A template for typical machine learning projects using Vertex AI.
flowchart LR
scheduler(Cloud Scheduler) -- HTTP Request --> functions(Cloud Functions)
functions --> vertex(Vertex AI Pipelines)
registry(Artifact Registry) -. Pipeline YAML .-> vertex
registry -. Dcoker Image .-> vertex
- Poetry
- Google Cloud SDK
- Terraform
poetry install
poetry shell
Create Google Cloud Platform resources with Terraform:
inv terraform.init -f invoke-dev.yaml
inv terraform.apply -f invoke-dev.yaml
Build Docker images for Vertex AI Pipelines components:
inv docker.build -f invoke-dev.yaml
Push Docker images:
inv docker.push -f invoke-dev.yaml
Generate pipeline YAML with Kubeflow Pipelines SDK:
inv pipeline.build -f invoke-dev.yaml
Upload pipeline YAML to Google Cloud Artifact Registry:
inv pipeline.push -f invoke-dev.yaml