Tuesday, 26 January 2016

Microsoft Makes Its CNTK Deep Learning Software More Accessible To Developers By Posting It On GitHub

Microsoft, GitHub to host the leading open-source computational network toolkit from CodePlex (CNTK) many developers to make it more accessible announced.

CNTK However, it was still on CodePlex, the scholars out of use was virtually unused, which means that an educational license, was limited by the one that became available in April 2015. Open Toolkit is a source of deep learning. On GitHub project by uploading, Microsoft action will be able to get more developers.

"This community open source code more quickly through the exchange of work to share ideas that will take advantage of CNTK is our hope," GitHub reads on the Microsoft site.

Performance-wise, CNTK speech and image recognition and other applications to make deep learning model is used in four other computational toolkits outdid. Microsoft addressed in Xuedong Huang, chief scientist of research and development, the toolkit is now more capable communications capabilities.

"CNTK toolkit we've ever seen anything more than just highly effective," Huang says.

In internal tests, CNTK Theano, torch 7, caf├ęs and TensorFlow, recently made clear that Google's machine learning system was compared with the likes of.

In December 2015 a group of researchers in the quality of the results of a more detailed document [PDF] published.

Microsoft, which is involved with the company's principal development manager Chris Basoglu, according to ", GPU-based computer large cluster" it can run on a single core, or the reasons why the upper hand against others says a toolkit. More to the point, researchers found that when it comes to large-scale experiments or computations which provides a huge advantage "in other publicly available toolkits, and more widely across GPU- based machines can "says.

Team GPU, images and movements as well as listen to acknowledge and understand the technology used in speech processing algorithms that effectively discovered. As such, to run Microsoft internal CNTK uses powerful computers with impressive GPUs. For example, the company's popular virtual assistant Cortana takes advantage of CNTK for speech recognition.

"We have more computational network toolkit (CNTK), GPU and CPU supports the introduction of an implementation of CN. We CNTK architecture and key components, CNTK to use the command-line options, and modification of the network definition and model description language, and acoustic model, language model, the language spoken and the setups provide a model for understanding, "the team in a publication [PDF] says.

Developers want.

"We did it to make changes because we want to give our customers the flexibility. That ecosystem of strong and toolset," Huang tells VentureBeat.

No comments:

Post a Comment

Note: only a member of this blog may post a comment.