MambaLiteSR: Image Super-Resolution with Low-Rank Mamba using Knowledge Distillation

TL;DR

MambaLiteSR is a lightweight, efficient image super-resolution model optimized for edge devices, utilizing knowledge distillation and low-rank approximation to reduce power consumption and parameters while maintaining high image quality.

Contribution

The paper introduces MambaLiteSR, a novel lightweight SR model that combines Vision Mamba architecture with knowledge distillation and low-rank approximation for resource-efficient deployment.

Findings

Outperforms state-of-the-art edge SR methods in power efficiency

Uses 15% fewer parameters with minimal performance loss

Achieves up to 58% power reduction on embedded devices

Abstract

Generative Artificial Intelligence (AI) has gained significant attention in recent years, revolutionizing various applications across industries. Among these, advanced vision models for image super-resolution are in high demand, particularly for deployment on edge devices where real-time processing is crucial. However, deploying such models on edge devices is challenging due to limited computing power and memory. In this paper, we present MambaLiteSR, a novel lightweight image Super-Resolution (SR) model that utilizes the architecture of Vision Mamba. It integrates State Space Blocks and a reconstruction module for efficient feature extraction. To optimize efficiency without affecting performance, MambaLiteSR employs knowledge distillation to transfer key insights from a larger Mamba-based teacher model to a smaller student model via hyperparameter tuning. Through mathematical analysis of model parameters and their impact on PSNR, we identify key factors and adjust them accordingly. Our comprehensive evaluation shows that MambaLiteSR outperforms state-of-the-art edge SR methods by reducing power consumption while maintaining competitive PSNR and SSIM scores across benchmark datasets. It also reduces power usage during training via low-rank approximation. Moreover, MambaLiteSR reduces parameters with minimal performance loss, enabling efficient deployment of generative AI models on resource-constrained devices. Deployment on the embedded NVIDIA Jetson Orin Nano confirms the superior balance of MambaLiteSR size, latency, and efficiency. Experiments show that MambaLiteSR achieves performance comparable to both the baseline and other edge models while using 15% fewer parameters. It also improves power consumption by up to 58% compared to state-of-the-art SR edge models, all while maintaining low energy use during training.