Efficient Hardware Realization of Convolutional Neural Networks using Intra-Kernel Regular Pruning

TL;DR

This paper introduces an intra-kernel regular pruning method for CNNs that significantly reduces parameters and computation while maintaining high accuracy, facilitating efficient hardware deployment.

Contribution

The paper presents a novel intra-kernel regular pruning scheme that preserves kernel structure, enabling hardware efficiency and substantial model compression.

Findings

Up to 10x parameter reduction

Up to 7x computational reduction

Less than 1% accuracy degradation

Abstract

The recent trend toward increasingly deep convolutional neural networks (CNNs) leads to a higher demand of computational power and memory storage. Consequently, the deployment of CNNs in hardware has become more challenging. In this paper, we propose an Intra-Kernel Regular (IKR) pruning scheme to reduce the size and computational complexity of the CNNs by removing redundant weights at a fine-grained level. Unlike other pruning methods such as Fine-Grained pruning, IKR pruning maintains regular kernel structures that are exploitable in a hardware accelerator. Experimental results demonstrate up to 10x parameter reduction and 7x computational reduction at a cost of less than 1% degradation in accuracy versus the un-pruned case.