Google renames TensorFlow Lite to LiteRT, TensorFlow brand remains

TensorFlow Lite (TFLite) was announced in 2017 and Google is now calling it “LiteRT” to reflect how it supports third-party models. 

TensorFlow Lite for mobile on-device AI has “grown beyond its TensorFlow roots to support models authored in PyTorch, JAX, and Keras.” 

Google says the new branding “captures this multi-framework vision for the future: enabling developers to start with any popular framework and run their model on-device with exceptional performance.”

Google is actively developing the runtime, as well as conversion and optimization tools, under the new name. It says to “look out for more updates coming to LiteRT, improving how you deploy classic ML models, LLMs, and diffusion models with GPU and NPU acceleration across platforms.” Moving forward, LiteRT is part of the Google AI Edge suite for ML on Android, iOS, and embedded devices. 

Meanwhile, the “main TensorFlow brand will not be affected.

This rebrand is rolling out progressively, and you’ll start seeing “LiteRT” in developer documentation, with tensorflow.org/lite redirecting to ai.google.dev/edge/litert.

Android developers accessing TensorFlow Lite via Google Play Services don’t have to make changes “at this time.”

• “If you currently use TensorFlow Lite via packages, you’ll need to update any dependencies to use the new LiteRT from Maven, PyPi, Cocoapods.”
• “If you currently build TensorFlow Lite from source, please continue building from the TensorFlow repo until code has been fully moved to the new LiteRT repo later this year.”