Conclusion

In this lesson, you covered a lot of ground. You explored converting a PyTorch model to a CoreML model, which is better suited for running locally on an iOS device. You then performed the same operation using scripts provided by the model developer. Afterward, you learned about the trade-offs that large models create on a mobile device where resources are limited compared to servers. You then explored compression models using CoreML Tools both directly and using the vendor’s script.

You’ve done a lot of work with little to show for it other than files. That’s about to change, as in the next lesson, you’ll integrate some of the models you’ve generated in this lesson into an iOS app using the Vision framework.