TensorFlow Lite, now named LiteRT, is still the same high-performance runtime for on-device AI, but with an expanded vision to support models authored in PyTorch, JAX, and Keras.
Overview
LiteRT, formerly known as TensorFlow Lite, is a high-performance runtime for on-device AI that now supports models from multiple frameworks including PyTorch, JAX, and Keras. This article discusses the rebranding, the transition process for developers, and the future vision for LiteRT.
What You'll Learn
1
How to transition from TensorFlow Lite to LiteRT with minimal code changes
2
Why LiteRT supports multiple frameworks for on-device AI
3
When to update dependencies for LiteRT in your applications
Key Questions Answered
What is the new name for TensorFlow Lite?
The new name for TensorFlow Lite is LiteRT, which reflects its expanded vision to support models from multiple frameworks like PyTorch, JAX, and Keras while maintaining high performance for on-device AI.
How can developers access LiteRT?
Developers can access LiteRT by updating their dependencies from repositories such as Maven, PyPi, and Cocoapods. For those using Google Play Services, no changes are necessary at this time.
Will existing TensorFlow Lite applications be affected by the name change?
No, existing applications using TensorFlow Lite will not be affected by the name change to LiteRT. They will continue to function as before without requiring any modifications.
What changes are expected in LiteRT's future development?
Future developments for LiteRT will focus on improving the deployment of classic ML models, LLMs, and diffusion models with GPU and NPU acceleration across various platforms.
Key Statistics & Figures
Number of apps using TensorFlow Lite
over 100K
These applications are running on approximately 2.7 billion devices, showcasing the widespread adoption of the technology.
Technologies & Tools
Some links below are affiliate links. We may earn a commission if you make a purchase.
Runtime
Litert
Used for deploying ML and AI models on Android, iOS, and embedded devices.
Framework
Pytorch
Supported for model conversion and deployment in LiteRT.
Framework
Jax
Supported for model conversion and deployment in LiteRT.
Framework
Keras
Supported for model conversion and deployment in LiteRT.
Key Actionable Insights
1Developers should update their dependencies to LiteRT to ensure they are using the latest features and optimizations.This is crucial as LiteRT will provide enhanced support for multiple frameworks, improving the performance and capabilities of on-device AI applications.
2Monitor the transition to LiteRT documentation for updates and best practices.Staying informed will help developers adapt quickly to changes and leverage new functionalities as they become available.
3Utilize the robust model conversion and optimization tools provided by Google AI Edge.These tools will facilitate the preparation of both open-source and custom models for deployment on various devices, enhancing the overall development workflow.
Common Pitfalls
1
Failing to update dependencies when transitioning to LiteRT can lead to compatibility issues.
Developers must ensure they are using the latest versions of LiteRT to take advantage of new features and optimizations.