Qualcomm Technologies Inc. a subsidiary of Qualcomm Incorporated has announced a new Snapdragon Neural Processing Engine SDK that accelerates deep learning on Snapdragon 820 series processors.
The announcement was made at Vision Summit in Santa Clara, California, where company revealed the SDK. The Snapdragon Neural Processing Engine SDK is powered by Qualcomm Zeroth Machine Intelligence platform and is highly optimized Snapdragon’s heterogeneous computing capabilities.
Using this OEM’s can provide deep learning-driven experiences on devices and make them more energy efficient. Zeroth technology is already in used by visual intelligence software like Snapdragon Scene Detect and advanced malware detection software seen in Snapdragon smart protect.
The Snapdragon Neural Processing Engine SDK is lightweight, flexible accelerated runtime for on-device execution of convolutional and recurrent neural networks on the Snapdragon 820 cores. It also supports common deep learning model frameworks, including Caffe and CudaConvNet. This new SDK will benefit range of companies in variety of industries like Healthcare, Automative, Security, Imaging etc.
It will be available in second half of 2016.
Gary Brotman, director of product management, Qualcomm Technologies, Inc. Commenting on the launch said – “The demand for untethered, mobile learning-driven user experiences is increasing rapidly and Qualcomm Technologies’ customers have been clamoring for tools to help them realize their product visions, With the introduction of the new Snapdragon Neural Processing Engine SDK, we are making it possible for myriad sectors, including mobile, IoT and automotive to harnesses the power of Qualcomm Snapdragon 820 and make high-performance, power efficient on-device deep learning a reality.”