Improve PCB inspection models toward your target.
What it does
Improves PCB inspection models through repeated evaluation, defect-data generation, mining, and retraining against an approved quality target.
When to use it
Use for a full AOI, PCB-defect, VisualChangeNet, or ChangeNet improvement workflow, not standalone training, inference, anomaly generation, or gap analysis.
How to use it
Give it an AOI model workspace and an iteration limit or time budget; after pre-flight approval, it runs the improvement loop and produces a model handoff and HTML report.
What you provide
- max_iterations or a time limit
- An existing project
Uses
Good to know
- acts without asking
What you need · 12
Requires Docker for container execution.
Requires NVIDIA Container Toolkit alongside Docker.
Requires an installed host Python that passes the bundled dependency probe; the launcher does not install an interpreter or dependencies.
Must be importable in the selected host Python.
Access · 0
No additional actions listed.