Neural Computing on a stick: deep learning hardware becomes affordable

Neural Computing on a stick: deep learning hardware becomes affordable

New Products |
By Christoph Hammerschmidt

As well as for use by data scientists and academics, the device is said to be ideal for a wide selection of developers and engineers looking at integrating computer-vision and artificial intelligence (AI) capabilities into their Internet of Things (IoT) applications. Available in the convenient USB form factor, the stick is enabling deep-learning prototyping on a laptop or any other computing device with a USB interface.

Like its predecessor– the Intel Movidius Neural Compute Stick launched in 2017 – the NCS 2 is based on the Intel Movidius Vision Processing Unit (VPU). However, it deploys the latest version: the Intel Movidius Myriad X VPU which has a dedicated hardware accelerator for DNN inference.

The NCS 2 also becomes a highly versatile prototyping and development tool when combined with the Intel Distribution of OpenVINO toolkit, which offers support for deep learning, computer vision and hardware acceleration to enable applications that have human-like vision capabilities. The combination of the NCS 2 and the Intel Distribution of OpenVINO toolkit enables a fast development-to-deployment cycle: from prototyping trained DNNs on the compute stick to the easy porting of DNNs to an Movidius VPU-based embedded device or system with minimal code changes required, if any. The NCS 2 also provides support for the popular open-source Caffe and TensorFlow DNN software libraries.

The Intel Neural Compute Stick 2 is shipping now from RS Components in the EMEA and Asia Pacific regions.

More information:

Related articles:

Intel brings Deep Learning to the engineering masses

Intel-powered UAV quadcopter platform targets educational markets

If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News


Linked Articles