Step 5 - SDK Packaging

In this step you will package your training run into a SDK package.

Required Steps

• If you are satisfied with the results, package the training run. You can either do this by clicking on Package button in the Training>Training view. Or, you can use the SDK Packaging>Packaging section of the platform in order to create a new package.

• You have now the first SDK package. You can test it using the Preview Functionality in the Mobile AI Data App (without doing any coding yet) or you can already do the final SDK integration.

Optional Steps

• If you are testing the results of your classification SDK and are not yet happy with the performance, try to understand by testing the SDK in real life, which labels are still not recognized sufficiently, or what exactly you want to improve. You can also take a closer look at the dataset that you packaged into the SDK package to see if this gives you any information on how to improve your model.

• If you have trained a Few Shot Classifier, you can try to finetune the small set of images for each label to better represent the label, or you could try to create a Classic Classifier model in your project and collect more data, if the recognition context is too complex for a Few Shot Classifier.

If at any point you need help from the Passio team, please reach out to us at [e-mail]