ReaLPrune: ReRAM Crossbar-aware Lottery Ticket Pruned CNNs

TL;DR

ReaLPrune introduces a crossbar-aware pruning method for CNNs on ReRAM architectures, enabling over 90% weight pruning, significant hardware reduction, and faster training without accuracy loss, suitable for edge ML applications.

Contribution

The paper presents a novel crossbar-aware pruning strategy that significantly improves hardware efficiency and training speed for CNNs on ReRAM-based edge devices, outperforming existing methods.

Findings

Prunes over 90% of CNN weights without accuracy loss.

Reduces hardware requirements by 77.2%.

Speeds up CNN training by approximately 20 times.

Abstract

Training machine learning (ML) models at the edge (on-chip training on end user devices) can address many pressing challenges including data privacy/security, increase the accessibility of ML applications to different parts of the world by reducing the dependence on the communication fabric and the cloud infrastructure, and meet the real-time requirements of AR/VR applications. However, existing edge platforms do not have sufficient computing capabilities to support complex ML tasks such as training large CNNs. ReRAM-based architectures offer high-performance yet energy efficient computing platforms for on-chip CNN training/inferencing. However, ReRAM-based architectures are not scalable with the size of the CNN. Larger CNNs have more weights, which requires more ReRAM cells that cannot be integrated in a single chip. Moreover, training larger CNNs on-chip will require higher power, which cannot be afforded by these smaller devices. Pruning is an effective way to solve this problem. However, existing pruning techniques are either targeted for inferencing only, or they are not crossbar-aware. This leads to sub-optimal hardware savings and performance benefits for CNN training on ReRAM-based architectures. In this paper, we address this problem by proposing a novel crossbar-aware pruning strategy, referred as ReaLPrune, which can prune more than 90% of CNN weights. The pruned model can be trained from scratch without any accuracy loss. Experimental results indicate that ReaLPrune reduces hardware requirements by 77.2% and accelerates CNN training by ~20X compared to unpruned CNNs. ReaLPrune also outperforms other crossbar-aware pruning techniques in terms of both performance and hardware savings. In addition, ReaLPrune is equally effective for diverse datasets and more complex CNNs