Performance Metrics

Page Outline

• Endpoints

---

Once training is complete, Model Studio generates a comprehensive suite of evaluation metrics. This category provides endpoints to retrieve specialized charts and reports like ROC curves, confusion matrices, and precision-recall curves to validate your model's readiness.

Endpoints

Endpoint	Method	Path	Description
Summary Metrics	GET	/{model_id}/evaluation/summary	Combined scorecards (Accuracy, F1, etc).
Confusion Matrix	GET	/{model_id}/evaluation/confusion-matrix	Accuracy breakdown per class.
ROC Curve	GET	/{model_id}/evaluation/roc-curve	TPR vs FPR over probability thresholds.
PR Curve	GET	/{model_id}/evaluation/pr-curve	Precision vs Recall trade-offs.
Classification Report	GET	/{model_id}/evaluation/report	Raw precision/recall/f1 per label.
Misclassification Errors	GET	/{model_id}/evaluation/errors	Samples where the model struggled most.