Google on Monday outlined plans to open source its machine learning system.
The system, called TensorFlow, is used in Google applications such as Google Photos and Google Translate as well as features such as smart reply and search.
Google uses TensorFlow to train its neural networks faster and improve products, but is looking for a developer boost. By open-sourcing TensorFlow, Google is looking to make its machine learning system more widely adopted.
We hope this will let the machine learning community--everyone from academic researchers, to engineers, to hobbyists--exchange ideas much more quickly, through working code rather than just research papers. And that, in turn, will accelerate research on machine learning, in the end making technology work better for everyone. Bonus: TensorFlow is for more than just machine learning. It may be useful wherever researchers are trying to make sense of very complex data--everything from protein folding to crunching astronomy data.
The move to open source TensorFlow comes as Facebook, Microsoft and other rivals are working to advance machine learning. Facebook took its machine learning and artificial intelligence tools to the open source community in January.
Among the key points:
- This open source release of TensorFlow supports individual computers and smartphones.
- TensorFlow is designed to run on CPUs, GPUs, desktops, servers and mobile computing platforms.
- The software is released under the Apache 2.0 license.
- Google has released a series of tutorials and data flow graphs to highlight how the system works. A data flow graph would look like this: