Event
VB Transform 2023 On-Demand
Did you miss a session from VB Transform 2023? Register to access the on-demand library for all of our featured sessions.
Register Now
Kneron, a leading artificial intelligence chip and edge computing provider, has announced the latest upgrade to its Neural Processing Unit (NPU). The NPU Upgrade is designed to help mitigate current GPU shortages caused by the demand growth of AI applications in the industry.
The NPU Upgrade is designed to offer similar performance to GPUs but without the increased heat dissipation, power consumption levels, and larger core size of such hardware. It will provide developers with faster access to features like deep learning, computer vision, image recognition, object identification, facial recognition, voice recognition, and anomaly detection.
The enhanced NPU offers greater optimization for edge and mobile-based AI applications. It includes support for TensorFlow, ONNX, PyTorch, and Caffe frameworks, as well as the OpenVX library. The NPU is energy efficient, with a 118-core arithmetic logic unit (ALU) which can reduce the number of calculations necessary for deep neural networks (DNN) processing, all without sacrificing accuracy.
Kneron claims that its NPU can offer up to 8 times the performance per watt of other implementations, with the ability to support more memory-intensive applications than traditional GPUs. It also supports a wide range of applications, from consumer devices to industrial machines.
The company’s CEO, Albert Liu, said: “Our advanced NPU Upgrade is tailored for those who don’t want to be limited by GPU availability. Our goal is to provide customers with an AI solutions that run faster, hotter and more efficiently at the edge.”
The NPU Upgrade is set to be available to market in the coming months. It will provide developers and AI teams with access to a powerful AI solution without the need to source expensive and scarce GPUs from the market.
