Movidius nerve calculation rod
This neuroplatics is the leading neural network developed by Movidius® Intel.
The prototype design and the deep neural network can be simply performed via USB rods, and the machine intelligence is operated with low power consumption. No cloud connection is required.
The USB size helps to easily connect to the existing host, while the visual processing unit (VPU) can reduce the machine load and accelerator intelligence applications.
Characteristics
Support convolutional neural network (CNN), such as caffe drawing, prototype design, and adjustment workflow
Provide all data and power through a single USB A port
In real time, in the device - no need to connect
Run multiple devices on the same platform to extend performance
Quickly deploy existing CNN models or unique networks
Movidius VPU adopts energy saving CNN process
Intel launchedMovidius nerve calculation rodThis is a USB mode-based deep learning reasoning tool and independentartificial intelligence(AI)The accelerator provides a dedicated depth neural network processing function for a wide range of edge host devices.ProficiencyMovidius?Neuroperative rods are designed for product developers, researchers and Cavern, providing dedicated high-performance depth neural network processing performance, thereby reducing obstacles to development, tuning and deploying artificial intelligence applications.
As more developers have adopted advanced machine learning solutions to develop and innovative applications and solutions, Intel is committed to providing comprehensive development tools and resources to ensure that developers reorganize in the era of digital economy with artificial intelligence as the core, equipment. Whether inIntel®Nervana? CloudTrain artificial intelligence neural networks, and via Intel®Strong®Expansion processors are artificial intelligence and virtual reality(VR)And augmented reality(Ar)And driverless(AD)Optimize the emerging field, or through the Movidius visual processing unit(VPU)Technology brings artificial intelligence to the edge, and Intel can provide a set of artificial intelligence tools, training and deployment options for the next generation of products and services that integrate into artificial intelligence.
Movidius(Intel''''''''s company)Vice President and General ManagerRemi El-Ouazzane said: "MOVIDIUS? The built -in MYRIAD 2 VPU built -in MYRIAD 2 VPU of the nerve computing rod provides strong and efficient performance -it can provide more than 100 billion floating -point operations per second at a power of 1 watt -in order to directly run real -time deep neural networks on the device(DNN). This makes various artificial intelligence applications deploy offline."
The development of machine intelligence mainly includes two phases:
? Through modern machine learning technology, the algorithm is trained with a large amount of sampling data;
? Run algorithms in terminal applications that need to interpret real world data. Among them, the second stage is also called "reasoning", which is reasonable on the edge(Local reasoning in equipment)Bring many benefits in reducing delay, power consumption, and privacy.
>>Compilation:Automatically put a trained, Caffe-based convolutional neural network(CNN)Convert into an embedded neural network that can be run on the MOVIDIUS Myriad 2 VPU by optimization.
>> Tuning:Active performance indicators for industry standards and custom neural networks achieve high efficiency tuning under low power consumption, thereby achieving good performance in real world. Verify scripts allow developers to compare the accuracy of the optimized model on the device and the original PC model.
>>accelerate:Movidius?The function of the nerve calculation rod can be used as a separate neural network accelerator, and can improve performance and energy efficiency only in the existing computing platform to add special depth learning reason.
References of Raspberry Pispberry-Pi Refer to:
https://ncsforum.movidius.com/discussion/118/movidius-nc-sdk-1-07-07-with-raspberry-pi-support
Development data download link:
https://developer.movidius.com/getting-started
Technical Consultation Forum:
https://ncsforum.movidius.com/discussions