Run Your First Evaluation

This path uses a provided YOLO Object Detection Dataset and trained .pt Model. You will import both into a new Project, generate predictions for clean images and one Fog severity, run a Robustness Evaluation, inspect its Claim and Evidence, and export a PDF Report.

What you will produce

By the end, your Project will contain:

• an independently owned Dataset Material;
• a configured and verified local Model;
• a sealed Prediction Set with clean and transformed predictions;
• a completed Evaluation with an immutable Claim and Evidence; and
• a Robustness PDF Report saved outside the Project where you choose.

The Evaluation does not need to Pass for this tutorial to succeed. A completed Evaluation with a valid Claim is a successful workflow; Pass, Fail, and N/A are scientific decisions about individual results.

Before you begin

• Use a supported Windows x64 computer.
• Obtain an active Knotest License that includes Evaluation and report export.
• Reserve enough time for local Model inference on the tutorial Dataset.
• Keep Knotest open while Prediction Generation and the Evaluation are running.

Continue to Install Knotest.