Note: Both TensorFlow Lite Metadata and the Android wrapper code generator are in experimental (beta) phase.
TensorFlow Lite metadata provides a structured framework for storing metadata to convey information for both the developer that will utilitised the model and code generators which can create wrapper around the model. For information on how to populate model metadata, please refer to the TensorFlow Lite Metadata documentation.
The first code generator which takes advantage of this metadata format is the TensorFlow Lite Android Code Generator. For more information on how to use this generator, please refer to the TensorFlow Lite Android wrapper code generator documentation.