Google is providing a software development kit, called ML Kit, that offers the company’s machine learning technologies to developers building Android and iOS mobile apps.
Featuring a set of base APIs to build machine learning into apps, ML Kit is now in public beta. It works with the Firebase mobile development platform. (Apple has its own Core ML machine learning kit for only iOS.)
ML Kit’s base APIs cover:
- Barcode scanning, to scan and process barcodes.
- Text recognition.
- Face detection.
- Landmark detection, to identify popular landmarks.
- Image labeling, to identify objects, locations, activities, products, and animal species.
Due soon is a smart reply API, providing suggested text snippets to fit a context.
ML Kit acts as an API layer to custom models to make it easy to use these models. Developers can use their existing TensorFlow Lite models with ML Kit. By being linked to Firebase, developers can experiment with different machine learning models using A/B testing. Google is also releasing an experimental model compression flow to reduce model sizes.
Where to download Google ML Kit
You can download Google ML Kit through Firebase.