Last year Google opened TensorFlow to the entire world. This meant that any individual, company, or organization could build their own AI applications using the same software that Google does to fuel everything from photo recognition to automated email replies. But there was a catch.
While Google stretches its platform across thousands of computer servers, the version it released to the public could run only on a single machine. This made TensorFlow considerably less capable for others. Google is trying to fix that deficiency now.
Today the company is releasing a new version of TensorFlow, and the most notable new feature is the ability to run it on multiple machines at the same time. Not everyone needs to run TensorFlow on hundreds, let alone thousands, of servers. But many researchers and startups could will benefit from being able to run TensorFlow on multiple machines.
The TensorFlow team opted to release a more limited version last year just to get something into researchers’ hands while continuing to work on more advanced features.
TensorFlow is based on a branch of AI called deep learning, which draws inspiration from the way that human brain cells communicate with each other. Deep learning has become central to the machine learning efforts of other tech giants such as Facebook, Microsoft, and Yahoo, all of which are busy releasing their own AI projects into the wild.
Early last year Facebook released some of the tools it uses to run the open source AI framework Torch across multiple servers. This year Microsoft open sourced its own AI framework that can run on multiple servers, and Yahoo quickly followed suit.
Despite the available alternatives, TensorFlow is already surprisingly popular. It was among the six open source projects that received the most attention from developers in all of 2015, even though it was only released in November. But it’s only now that TensorFlow has been unshackled from the one-machine limit that we’ll start to really see what it’s capable of.