CompressNAS : A Fast and Efficient Technique for Model Compression using Decomposition
This addresses the challenge of model compression for microcontrollers and lightweight NPUs, offering an incremental improvement over existing local rank selection methods.
The paper tackles the problem of deploying deep CNNs on resource-constrained devices by introducing CompressNAS, a framework for global rank selection in tensor decomposition, achieving up to 8x compression on ResNet-18 with less than 4% accuracy drop on ImageNet and 2x compression on YOLOv5 models with minimal accuracy loss on COCO.
Deep Convolutional Neural Networks (CNNs) are increasingly difficult to deploy on microcontrollers (MCUs) and lightweight NPUs (Neural Processing Units) due to their growing size and compute demands. Low-rank tensor decomposition, such as Tucker factorization, is a promising way to reduce parameters and operations with reasonable accuracy loss. However, existing approaches select ranks locally and often ignore global trade-offs between compression and accuracy. We introduce CompressNAS, a MicroNAS-inspired framework that treats rank selection as a global search problem. CompressNAS employs a fast accuracy estimator to evaluate candidate decompositions, enabling efficient yet exhaustive rank exploration under memory and accuracy constraints. In ImageNet, CompressNAS compresses ResNet-18 by 8x with less than 4% accuracy drop; on COCO, we achieve 2x compression of YOLOv5s without any accuracy drop and 2x compression of YOLOv5n with a 2.5% drop. Finally, we present a new family of compressed models, STResNet, with competitive performance compared to other efficient models.