Google Research opens source for its machine learning framework


Google has open sourced TensorFlow, a key algorithm that helps it with voice recognition, image search with Google Photos and several projects in automated image captioning.

According to a post in its official blog, Jeff Dean, Senior Google Fellow, and Rajat Monga, Technical Lead at Google Research said “we’re proud to announce the open source release of TensorFlow – our second-generation machine learning system, specifically designed to correct these shortcomings.”

According to Google Research, TensorFlow is general, flexible, portable, easy-to-use, and completely open source. The blogpost added that this move has happened while improving upon DistBelief’s speed, scalability, and production readiness. The post added that according to some benchmarks, TensorFlow is twice as fast as DistBelief.

According to a post in the Christian Science Monitor, a neural network allows the Google Photos app to learn the relationship between an object’s name and its appearance. This allows Google Photos to identify a cat within a photo based on similar pictures it’s seen before. Similarly, another neural network allows the Google Translate app to learn how words are commonly used in conversation, so it can provide less-stilted renderings.

The Google Research blog post added that TensorFlow is ready for use in real products too. It added that TensorFlow was built from the ground up to be fast, portable, and ready for production service and can help you move your idea seamlessly from training to your desktop GPU and even running it on your mobile phone.

TensorFlow also lets you get started quickly with powerful machine learning tech by using Google’s example model architectures.