Use custom models with the SDK to monitor usage and performance, or deploy updated models over-the-air. Support for iOS Core ML, TensorFlow Lite and TensorFlow mobile.
Custom model tracking allows you to monitor usage and performance on all devices running your app, live in production.
Push model version updates over-the-air, no application update needed, no waiting for the app store approval.
Sample input/output data from real-world devices, retrain models, catch edge cases, and improve accuracy.
One SDK supports all of the widely-used machine learning runtimes. Use Core ML, TensorFlow Lite, and TensorFlow mobile.
Creating your edge-deployed models does not need to create extra risk. Learn more about on-device model protection to avoid model theft and inspection.