top

Google releases developer preview of TensorFlow Lite

Back in the month of June, Google made an announcement at Google I/O about the new version of TensorFlow. Finally on Tuesday, 14th November, Google released the developer preview of TensorFlow Lite. The search giant has released the software library with the aim of creating more lightweight machine learning solutions for smartphones and embedded devices. Today more and more mobile devices are integrated with purpose-built custom hardware in order to process machine learning workloads efficiently. Google’s TensorFlow Lite supports the Android Neural Networks API that could help in quick initialization and improvement in model load times on a variety of mobile devices. The primary purpose of TensorFlow Lite is to bring low-latency inference from machine learning models to devices that are relatively less robust. To put it simply, rather than learning a new power from some existing data, TensorFlow Lite will aim at applying the existing power of models to the new data provided. As per the Google Official Developer’s Blog, “With this developer preview, we have intentionally started with a constrained platform to ensure performance on some of the most important common models. We plan to prioritize future functional expansion based on the needs of our users.” TensorFlow Lite has already got support for a number of models that have been trained and optimized for mobile. The natural language processing models include MobileNet, Inception v3, and Smart Reply.  
Rated 4.0/5 based on 20 customer reviews
Normal Mode Dark Mode

Google releases developer preview of TensorFlow Lite

Ruslan Bragin
What's New
17th Nov, 2017
Google releases developer preview of TensorFlow Lite

Back in the month of June, Google made an announcement at Google I/O about the new version of TensorFlow. Finally on Tuesday, 14th November, Google released the developer preview of TensorFlow Lite. The search giant has released the software library with the aim of creating more lightweight machine learning solutions for smartphones and embedded devices.

Today more and more mobile devices are integrated with purpose-built custom hardware in order to process machine learning workloads efficiently. Google’s TensorFlow Lite supports the Android Neural Networks API that could help in quick initialization and improvement in model load times on a variety of mobile devices.

The primary purpose of TensorFlow Lite is to bring low-latency inference from machine learning models to devices that are relatively less robust. To put it simply, rather than learning a new power from some existing data, TensorFlow Lite will aim at applying the existing power of models to the new data provided.

As per the Google Official Developer’s Blog, “With this developer preview, we have intentionally started with a constrained platform to ensure performance on some of the most important common models. We plan to prioritize future functional expansion based on the needs of our users.”

TensorFlow Lite has already got support for a number of models that have been trained and optimized for mobile. The natural language processing models include MobileNet, Inception v3, and Smart Reply.

 
Ruslan

Ruslan Bragin

Author
Ruslan is a passionate in developing data and Machine learning solution. He is currently working on projects related to IoT.

Leave a Reply

Your email address will not be published. Required fields are marked *

SUBSCRIBE OUR BLOG

Follow Us On

Share on

other Blogs

20% Discount