Reposted from the Microsoft Next blog.
Microsoft is releasing its Computational Network Toolkit (CNTK) on GitHub, making the very efficient AI tools used by its own researchers available to the broad developer and data science community.
Xuedong Huang, Microsoft’s Chief Speech Scientist, and his team were anxious to make faster improvements to how well computers can understand speech, and the tools they had to work with were slowing them down. So, a group of volunteers set out to solve this problem using a homegrown solution that stressed performance over all else.
CNTK is the outcome of that project, and has proved more efficient than other popular computational toolkits used by developers to create deep learning models for speech and image recognition. Microsoft is internally using CNTK on a set of powerful computers that use graphics processing units, or GPUs.
We are making these tools available to other researchers who are making similar advances in deep learning via open source licenses. We hope many other developers will join us as we drive artificial intelligence breakthroughs.
As Huang puts in, “The CNTK toolkit is just insanely more efficient than anything we have ever seen.”
ML Blog Team