Understanding all the parts of the Core ML SDKs is important before diving too deep. In this video, join Emmanuel Henri to get an introduction.
- [Instructor] Let's explore the architecture behind Core ML so you understand better what happens behind the scenes. Core ML's models run on a device, leveraging the CPU and GPU while minimizing its footprint. The advantage of running it on your device removes the need of having a connection with a server. It also provides a more secure environment for isolating the data used in the device versus sending this data to an external server where the data is more prone to security issues. The way it is structured on your device is as follows. Your application generates data on what it uses, for instance, images, texts, and activity data, which is then fed into the Core ML SDKs and models and where machine learning can be applied. The model inputs and generated outputs stay on your device and are isolated also inside of your application, so hopefully, this helps bring a bit more clarity into what happens behind the scenes with Core ML.
- The Core ML architecture
- Setting up a new project with the Create ML tool
- Testing a model with new data
- Integrating an ML model into an iOS app
- Converting models from other projects