PaPr: Training-Free One-Step Patch Pruning with Lightweight ConvNets for Faster Inference

arXiv:2403.16020v1 Announce Type: new
Abstract: As deep neural networks evolve from convolutional neural networks (ConvNets) to advanced vision transformers (ViTs), there is an increased need to eliminate redundant data for faster processing without compromising accuracy. Previous methods are often architecture-specific or necessitate re-training, restricting their applicability with frequent model updates. To solve this, we first introduce a novel property of lightweight ConvNets: their ability to identify key discriminative patch regions in images, irrespective of model's final accuracy or size. We …

abstract accuracy advanced advanced vision architecture arxiv convolutional neural networks cs.cv data faster free inference networks neural networks processing pruning training transformers type vision vision transformers