To accelerate the development of artificial intelligence and speech recognition projects, Microsoft researchers have developed a toolkit for deep neural networks and multi-GPU acceleration processing. But now, the company has further open sourced the ComputaTIonal Networks Toolkit (CNTK) and put the full content on GitHub. Microsoft claims that CNTK is faster than other publicly available deep learning tools, but the company is not the first technology giant to open deep learning software.
In November last year, Google opened up the TensorFlow machine learning system for its own image search function. But according to Microsoft's chief speech scientist, the efficiency of the company's CNTK toolkit is "crazy than we have ever seen."
This is partly due to the ability of CNTK to leverage the capabilities of a graphics processing unit (GPU), which Microsoft claims to be the only company that exposes "scalable GPU" functionality. (from one on a single machine to a number on a supercomputer)
After matching the company's networked GPU system (called Azure GPU Lab), it will be able to train deep neural networks to recognize speech, making Cortana virtual assistants ten times faster.
In April of last year, CNTK was already open to researchers, but there were quite a few restrictions on open source licensing at the time. But now, it's completely open, and the beginners of deep learning will benefit the most.
Epoxy Coated Thermistor
Epoxy coated NTC Thermistor is small size and fast response. It also has the properties of high precision testing protection, good consistency and interchangeability as well as long-term stable operation. This type of thermistor can be used in air conditioner, central heating, clinical thermometer, automotive electronics, electronic calendar and so on.
Epoxy Resin Coated Thermistor,Thermistor Digital Thermometer,Air Conditioner Sensor,Automotive Thermistor
Feyvan Electronics Technology Co., Ltd. , http://www.fv-cable-assembly.com