A framework has no prebuilt serving image. What do you use?
Why C is the answer
A custom container
Custom containers package any framework.
In detail
- A custom container serves any framework.
- Prebuilt images are simpler if they exist.
Use case
An illustrative scenario, not a real incident.
A team uses a rare framework, and packages it in a custom container for serving.
See it in different ways
flowchart TD
A{"Is a caller waiting?"} -->|No| B["Batch inference"]
A -->|Yes| C{"How much control?"}
C -->|"Managed"| D["Agent Platform endpoint"]
C -->|"Own containers"| E["Cloud Run or GKE"]mindmap
root((Serving))
Online
Autoscaled backend
Batch
Overnight tables
Private endpoint
Internal only
Edge
No connection
Custom container
Any frameworkflowchart TD
Q{"Is a prebuilt image available?"}
Q -->|"Yes"| A0["Use it"]
Q -->|"No"| A1["Custom container"]Key point: Custom containers package any framework.
Why the other options are wrong
- A. A Workbench instance
- A notebook instance does not serve models.
- B. A prebuilt scikit-learn image
- That image is for scikit-learn only.
- D. A BigQuery ML model
- BigQuery ML is a different route.