Model Compression

TL;DR

This paper explores model compression techniques, combining pruning and quantization, and proposes a metric to optimize the balance between accuracy loss and size reduction in machine learning models.

Contribution

It introduces a new quality measurement metric for evaluating the effectiveness of combined pruning and quantization methods.

Findings

Effective combination of pruning and quantization reduces model size significantly.

The proposed metric helps in selecting optimal compression strategies.

Compression maintains acceptable accuracy levels while reducing resource requirements.

Abstract

With time, machine learning models have increased in their scope, functionality and size. Consequently, the increased functionality and size of such models requires high-end hardware to both train and provide inference after the fact. This paper aims to explore the possibilities within the domain of model compression, discuss the efficiency of combining various levels of pruning and quantization, while proposing a quality measurement metric to objectively decide which combination is best in terms of minimizing the accuracy delta and maximizing the size reduction factor.