Microsoft has transformed its Computational Network Toolkit (CNTK) into an upgraded Cognitive Toolkit to let developers easily deploy deep learning technologies on their projects. The new release is particularly designed for speed advances in areas such as speech and image recognition search relevance on CPUs and GPUs.
Available as Cognitive Toolkit 2.0 beta, the framework emerges from a research tool that Microsoft and its core team used in the past for internal research processes.
“We have taken it from a research tool to something that works in a production setting,” said Frank Seide, a principal researcher at Microsoft Artificial Intelligence and Research and a key architect of Microsoft Cognitive Toolkit, in a blog post.
While the original CNTK was accessible through C++, the latest Cognitive Toolkit comes with native support for Python 3. It will also receive Python 2 support in the coming future to expand the coverage of the open source deep learning model.
The Microsoft team, managed by Seide, has also improved the performance of the framework to enable its presence on big datasets across multiple machines. This gives room for scalability across various high-end NVIDIA GPUs and machines on large datasets.
“Microsoft Cognitive Toolkit represents a tight collaboration between Microsoft and NVIDIA to bring advances to the deep learning community,” said Ian Buck, general manager of the Accelerated Computing Group at NVIDIA, in a statement. “Compared to the previous version, it delivers almost two times performance boost in scaling to eight Pascal GPUs in an NVIDIA DGX-1.”
Apart from the NVIDIA GPU support, the community-centric framework comes with Azure networking capabilities to provide deep learning support on technologies powered by the cloud.
It is deep learning all around
Microsoft is not the only technology giant to work on deep learning developments. Google and Facebook are also moving parallel with their TensorFlow and Torch offerings, respectively.
However, Microsoft is claiming to have already morphed into a full-fledged offering. It is presently used not just by some flagship products of the software giant but also available in some smart appliances.
Developers can certainly take Microsoft’s Cognitive Toolkit to new levels by customising its presence through a GitHub repository. Also, there is a type of artificial intelligence work called reinforcement learning to bring AI to the masses.
