Flexible and Fully Quantized Ultra-Lightweight TinyissimoYOLO for Ultra-Low-Power Edge Systems

TL;DR

This paper evaluates TinyissimoYOLO, a fully quantized ultra-lightweight object detection network, on various ultra-low-power edge platforms, demonstrating its efficiency and flexibility for real-time inference in power-constrained environments.

Contribution

It introduces and characterizes variants of TinyissimoYOLO optimized for ultra-low-power edge systems, and compares their performance across multiple state-of-the-art hardware platforms.

Findings

GAP9's hardware accelerator achieves 2.12ms inference latency and 150uJ energy consumption.

GAP9 can run higher resolution TinyissimoYOLO with 112x112 pixels in 3.2ms at 245uJ.

Multi-core GAP9 implementation achieves 11.3ms latency and 490uJ energy efficiency.

Abstract

This paper deploys and explores variants of TinyissimoYOLO, a highly flexible and fully quantized ultra-lightweight object detection network designed for edge systems with a power envelope of a few milliwatts. With experimental measurements, we present a comprehensive characterization of the network's detection performance, exploring the impact of various parameters, including input resolution, number of object classes, and hidden layer adjustments. We deploy variants of TinyissimoYOLO on state-of-the-art ultra-low-power extreme edge platforms, presenting an in-depth a comparison on latency, energy efficiency, and their ability to efficiently parallelize the workload. In particular, the paper presents a comparison between a novel parallel RISC-V processor (GAP9 from Greenwaves) with and without use of its on-chip hardware accelerator, an ARM Cortex-M7 core (STM32H7 from ST Microelectronics), two ARM Cortex-M4 cores (STM32L4 from STM and Apollo4b from Ambiq), and a multi-core platform with a CNN hardware accelerator (Analog Devices MAX78000). Experimental results show that the GAP9's hardware accelerator achieves the lowest inference latency and energy at 2.12ms and 150uJ respectively, which is around 2x faster and 20% more efficient than the next best platform, the MAX78000. The hardware accelerator of GAP9 can even run an increased resolution version of TinyissimoYOLO with 112x112 pixels and 10 detection classes within 3.2ms, consuming 245uJ. To showcase the competitiveness of a versatile general-purpose system we also deployed and profiled a multi-core implementation on GAP9 at different operating points, achieving 11.3ms with the lowest-latency and 490uJ with the most energy-efficient configuration. With this paper, we demonstrate the suitability and flexibility of TinyissimoYOLO on state-of-the-art detection datasets for real-time ultra-low-power edge inference.