Evaluation
Measure how well a trained scenario model performs by scoring it against held-out data with known outcomes, then turn those results into a readable report.
In This Section
| Page | Description |
|---|---|
| Testing Against Ground Truth | Run test() on the held-out test set with configurable metrics and output types, exporting predictions alongside ground truth. |
| Evaluation Report | Generate a Markdown report with task-specific metrics and diagnostic plots from a predictions file. |