CoreML is Apple's on-device machine learning inference framework, designed to run models efficiently on iPhones and other Apple hardware by routing computation through Apple's Neural Engine for lower latency and reduced power draw. Models are stored as .mlmodel files and can be created with Apple's CreateML tool or converted from TensorFlow .h5 or PyTorch .pt formats. For developers targeting the Apple ecosystem, CoreML is the standard path to deploying trained models directly on device without a network call.
"ML in a Minute" is our conversational series on answering machine learning questions. Have questions you want answered? Tweet at us.
What is CoreML (in 60 Seconds or Fewer)?
CoreML is Apple's machine learning framework for doing on device inference. When you're doing on device inference, you want to be especially considerate of creating a model that is small, low latency, and uses low power consumption. CoreML allows you to easily have a model file – known as a .mlmodel file in the Apple ecosystem– to deploy to iPhones and Apple devices. The model format is optimized to make use of Apple's Neural Engine, so that the models run faster, with lower power consumption.
You can create a CoreML model using some a tool like CreateML. You can also import your .h5 file from TensorFlow or your .pt file from Python. Basically, CoreML is an easy way to get up and going in the Apple ecosystem for on device machine learning.
Liked this? Be sure to also check out the computer vision glossary.
Cite this Post
Use the following entry to cite this post in your research:
Joseph Nelson. (Aug 16, 2021). What is CoreML?. Roboflow Blog: https://blog.roboflow.com/what-is-coreml/