SigmaQuant: Hardware-Aware Heterogeneous Quantization Method for Edge DNN Inference

TL;DR

SigmaQuant is an adaptive, hardware-aware heterogeneous quantization framework that optimizes DNN inference on edge devices by balancing accuracy and resource constraints without extensive search.

Contribution

It introduces SigmaQuant, a novel adaptive layer-wise quantization method that efficiently adapts to hardware conditions, improving over uniform and existing heterogeneous approaches.

Findings

Achieves better accuracy-resource trade-offs on edge devices.

Reduces the need for exhaustive search in quantization design.

Adapts to diverse hardware constraints effectively.

Abstract

Deep neural networks (DNNs) are essential for performing advanced tasks on edge or mobile devices, yet their deployment is often hindered by severe resource constraints, including limited memory, energy, and computational power. While uniform quantization provides a straightforward approach to compress model and reduce hardware requirement, it fails to fully leverage the varying robustness across layers, and often lead to accuracy degradation or suboptimal resource usage, particularly at low bitwidths. In contrast, heterogeneous quantization, which allocates different bitwidths to individual layers, can mitigate these drawbacks. Nonetheless, current heterogeneous quantization methods either needs huge brute-force design space search or lacks the adaptability to meet different hardware conditions, such as memory size, energy budget, and latency requirement. Filling these gaps, this work introduces \textbf{\textit{SigmaQuant}}, an adaptive layer-wise heterogeneous quantization framework designed to efficiently balance accuracy and resource usage for varied edge environments without exhaustive search.