AI processing technology to bring lower energy to automotive industries

At CES 2020 in Las Vegas, Blaize (South Hall 2, Booth #25332) will demonstrate the high performance ailia-based AI framework running on the Blaize GSP platform.

“Blaize and AXELL have worked closely throughout a six month test and qualification programme that proved flexible programming and superior performance of the Blaize GSP architecture in running the AXELL neural networks at the heart of the ailia framework,” said Kazuki Kyakuno, a technical board member of AXELL CORPORATION. “AXELL plans to go forward with end solutions that utilise the GSP breakthrough processing architecture, optimised for AI inferencing at the extreme edge to the core.”

“We are pleased to collaborate with AXELL, a leader in smart vision solutions, to bring breakthroughs in energy efficiency to AI enabled smart vision at the edge,” said Dinakar Munagala, CEO, Blaize.

Designed for executing neural networks at the edge, Blaize GSP based systems provide a more efficient way to run standard and customised neural nets in vehicles, retail stores and electronic game devices. Blaize GSP delivers a 10-100x systems efficiency advantage over legacy AI hardware solutions, coupled with a single scalable architecture across a range of AI use cases.

AXELL ailia SDK is a cross-platform high speed inference SDK. The ailia SDK provides a consistent C++ API on Windows, Mac, Linux, iOS and Android. It supports Unity, Python and JNI for efficient AI implementation. The ailia SDK makes great use of the GPU to serve accelerated computing. AXELL is pleased to start supporting embedded devices like those based on Blaize GSP.