Embedded Neural Network Compute Framework:

Deep Learning Deployment on Myriad™ 2

Thanks to its highly efficient mix of programmable compute and fixed hardware accelerators, Myriad™ 2 is able to deliver class-leading performance on deep neural networks. Running within a single Watt, Myriad 2 can run industry standard neural networks, or can even be programmed to run custom networks through commonly supported layer types. 

Key Benefits

Run neural networks at the edge -no cloud connection required
Ultra-Low Power performance
Combine with traditional imaging and vision algorithms
Run custom neural networks with commonly support layer types

Intelligent Deep Learning Framework

Myriad™ 2 comes with a software framework translates trained neural networks from a PC environment to an embedded environment. The software development kit intelligently optimizes Caffe networks to run on the ultra-low power Myriad 2 VPU.

Prototype, Validate, Compute

Included profiling tools enable developers to rapidly evaluate performance of a neural network on a layer-by-layer basis. Validation scripts allow developers to compare accuracy between PC and embedded models.

Breaking New Ground in Computing

Deep Neural Networks (DNNs) are solving some of the most challenging problems in modern computing. DNNs have been shown to drastically outperform traditional approaches on tasks such as image classification, voice recognition and complex problem solving. DNNs have the potential to make our devices more aware, more proactive and more useful than ever before.

The 12 SHAVE cores on Myriad 2 can be tasked with a mix of algorithms and vision tasks. Implement deep neural network based algorithms alongside traditional computer vision algorithms.

Related Videos