NPU integrates ISP for edge-AI

Author: EIS Release Date: Aug 23, 2023

Kneron, the Qualcomm-backed neural processing specialist, is about to start sampling an automotive-grade neuron processing unit (NPU) with an integrated image signal processor (ISP).

 

The KL730 has a peripheral interface that connects various digital signals like image, video, audio, and millimeter waves.

 

 

 

 

The device claims to deliver a 3 to 4X improvement in energy efficiency compared to previous Kneron devices,

 

 

Aimed at edge-AI,  the chip supports compute power from 0.35 to 4 effective TOPS, expanding its ability to support advanced lightweight GPT large language models like the nanoGPT.

 

The KL730 addresses security in the AIoT  by enabling users to run GPT models either partially or fully offline.

 

The chip utilizes Kneo, Kneron’s private  edge AI network, to permit AI to reside on users’ edge devices, offering them more control over data privacy.

 

For example, engineers can design ICs without sharing confidential data with large cloud companies operating data centers. .