LeapMind is an edge AI firm that aims to implement machine learning under the corporate philosophy of "To create innovative devices with machine learning and make them available everywhere". Our expertise lies in our extremely low bit quantization, which is a lightweight deep learning model, and we have a successful collaborative development record with more than 150 companies, mainly in the automobile and other manufacturing industries. Based on the expertise gained from the development of both software and hardware, we have developed and provided "Efficiera", Ultra-low power AI inference accelerator IP.
Efficiera is optimized for CNN inference processing on FPGAs and ASIC/ASSP devices, which require energy efficient and a small footprint, and is a solution to help integrate AI into edge devices. It consists of a lightweight deep learning model that has been reduced weight by extremely low bit quantization, AI accelerator IP cores running on FPGA/ASIC systems optimized for image recognition to run the model high efficiently, and other support. Efficiera provides excellent power and area efficiency, contributing to energy efficient and cost reduction of AI-based products, thereby accelerating the mass production of our clients' on-device AI solutions. Particularly in FPGAs, Efficiera's small footprint allows integration of Efficiera on the same SoC FPGA device as the CPU and image input circuitry, which helps to reduce the cost of (BoM) for products which comes with AI (Artificial Intelligence).