Tool for automatically designing deep learning models that run efficiently on R-Car

When designing the architectures of deep learning models for AD/ADAS applications, we often develop them based on deep learning models proposed in open source. However, those models do not always work with high-performance efficiency in R-Car and may not meet the performance requirements. In this case, a rework to the design phase of the deep learning model architecture will be required, and in some cases, many man-hours will be required to manually redesign the deep learning model architecture to run with high performance on the R-Car.

This tool automatically searches for a highly efficient deep learning model that takes into account the hardware characteristics of R-Car, thus enabling the development of a deep learning model that meets the performance requirements and is more accurate in a short period of time, without requiring in-depth knowledge of deep learning model design or R-Car hardware.

Target Devices

Design & Development

Tools to Optimize AI Software for AD/ADAS on R-Car SoC