After training your model, and you'll need to evaluate your model performance. Using Test Modelfunctionality, Plainsight provides an easy way to visually test your model's accuracy with new input by displaying detections on top of your image data.
Testing Your Model Version
Images for testing must be available to upload from your local computer. Supported image types are JPG, PNG, BMP, TIF, and WEBP. Max file size is 10 MB.
Testing the model with a novel image
The model version must be deployed in order to use the Test Model feature. Once the API is no longer active, the Test Model button will be disabled.
Navigate to the model version details of the model version you want to test. Under the Predictions API section, check that the Image API shows an Active status. If not, you will need to deploy the model by clicking "Start Image API".
At the bottom section of the dashboard, click the "Test Model" button on the far right.
Drag an image into the modal or click "choose a file" to select an image from your computer to test.
Your image will be processed and used as input for model inference against this version. When finished, your model detections be shown overlaid on your image.