Title :
The Computational Network Toolkit [Best of the Web]
Author :
Dong Yu ; Kaisheng Yao ; Yu Zhang
Author_Institution :
Miscrosoft Res., Bothell, WA, USA
Abstract :
The computational network toolkit (CNTK) is a general-purpose machine-learning tool that supports training and evaluation of arbitrary computational networks (CNs), i.e., machine-learning models that can be described as a series of computational steps. It runs under both Windows and Linux and on both central processing unit (CPU) and Compute Unified Device Architecture (CUDA)-enabled graphics processing unit (GPU) devices. The source code, periodic release builds, documents, and example setups can all be found at http://cntk.codeplex.com.
Keywords :
Linux; graphics processing units; learning (artificial intelligence); parallel architectures; recurrent neural nets; CNTK; CPU; CUDA-enabled GPU devices; Linux; Windows; arbitrary computational network evaluation; arbitrary computational network training; central processing unit; computational network toolkit; compute unified device architecture; general-purpose machine-learning tool; graphics processing unit; Computational modeling; Data models; Graphics processing units; Hidden Markov models; Neural networks; Speech recognition; Web sites;
Journal_Title :
Signal Processing Magazine, IEEE
DOI :
10.1109/MSP.2015.2462371