Integrating custom models using CoreML
Generally, ML models are trained to perform a specific task – recognizing a sentence’s sentiment, detecting humans, or analyzing sounds are all examples of different tasks done using various models. This means that even though the potential of the existing models is enormous, we are still limited in what we can do.
This is where the CoreML framework enters the picture. Using CoreML, we can integrate ML models that are not part of the iOS SDK, and we can even train our own models and add more intelligent capabilities.
It’s best to explain how to do this by using an example, such as detecting spam messages.
Imaging we are developing a messaging app. One of the most popular messaging app features is the ability to detect spam to improve the user experience and increase retention.
We must create an ML model to classify messages as spam to implement a spam detector.
To achieve this, we can use a desktop application...
