LiteRT
Run TensorFlow Lite models on phones and other edge devices with a lightweight inference runtime.
Published by Google AI Edge
How to use it
Bundle a supported TFLite model with the application and validate the expected input shape, preprocessing, and output interpretation on a real device.
Pick the runtime and acceleration path that match the deployment target. Benchmark CPU and available delegates on representative hardware because unsupported operators and device drivers can affect the result.
- Hardware
- CPU, Mobile GPU delegate
- Model formats
- TensorFlow Lite
- Operating systems
- Android, iOS, Linux
- Access
- Runs on device. Package the runtime and the TFLite model with the application.