Nvidia’s open source Deep Learning Accelerator (NVDLA) is to be integrated into Arm’s Project Trillium platform for machine learning. The two processing companies are collaborating to make it easier for Internet of Things (IoT) chip makers to integrate AI into their designs.
Mobile, consumer electronics and IoT are the target markets for this release, but many chip designers using vision technology could benefit from deep learning inferencing, i.e. applying what the neural network has learned in its training to data it’s never seen.
Machine learning techniques have been heralded as a tool that will open up large numbers of vision applications. In industrial vision, libraries like Halcon and Adaptive Vision’s software package are now incorporating machine learning functionality.
Arm's Project Trillium suite of processors are designed for machine learning; one chip is for object detection and can run real-time detection with full HD processing at 60fps.
‘Inferencing will become a core capability of every IoT device in the future,’ said Deepu Talla, vice president and general manager of autonomous machines at Nvidia. ‘Our partnership with Arm will help drive this wave of adoption by making it easy for hundreds of chip companies to incorporate deep learning technology.’
‘Accelerating AI at the edge is critical in enabling Arm’s vision of connecting a trillion IoT devices,’ said Rene Haas, executive vice president, and president of the IP group, at Arm. ‘We are one step closer to that vision by incorporating NVDLA into the Arm Project Trillium platform, as our entire ecosystem will immediately benefit from the expertise and capabilities our two companies bring in AI and IoT.’
Based on Nvidia’s Xavier chip, NVDLA is a free, open architecture to promote a standard way to design deep learning inference accelerators. NVDLA’s modular architecture is scalable, highly configurable and designed to simplify integration and portability.
NVDLA brings a host of benefits that speed the adoption of deep learning inference. It is supported by Nvidia’s suite of developer tools, including upcoming versions of TensorRT, a programmable deep learning accelerator. The open-source design allows for features to be added regularly, including contributions from the research community.
The integration of NVDLA with Project Trillium will give deep learning developers high levels of performance as they leverage Arm’s flexibility and scalability across the wide range of IoT devices.
Related article:
Teaching machines new tricks - Artificial intelligence seems to be sweeping the world and neural networks are now starting to find their way into the industrial imaging market. Greg Blackman investigates
