GaLe: memory-efficient Global Approximate and Local Exact features

2026-09-02

AI Summary

GaLe技术通过将特征图分为精确局部和近似全局两部分，解决了嵌入式设备上部署预训练网络的资源限制问题，实现了高效推理。

Abstract

Embedded devices typically lack the resources of GPU-equipped machines, and existing inference methods suffer from either high computational overhead (patch-based) or accuracy loss (approximation-based). We propose GaLe, a memory-efficient technique that enables the deployment of pretrained networks on constrained devices without retraining. GaLe partitions feature maps into two components: a local exact (Le) representation that preserves fine details and a global approximate (Ga) representation that retains long-range dependencies. Unlike standard tiling, GaLe supports global operations and attention mechanisms found in hybrid CNN-transformer models. Validated on ImageNet, our method matches exact-inference performance while achieving up to 65% speedup and 90% RAM reduction on a Cortex-M33 compared to patch-based inference. We further demonstrate GaLe's versatility across classification, detection, and generation tasks, highlighting its potential as a foundation for resource-efficient architecture design.

Problem

• memory-efficient
• pretrained networks
• constrained devices
• inference methods
• resource-efficient

Innovation

• memory-efficient
• Global Approximate and Local Exact features
• hybrid CNN-transformer models