Degree-Quant: Quantization-Aware Training for Graph Neural Networks

TL;DR

Degree-Quant introduces a novel quantization-aware training method for GNNs that maintains high accuracy with low-precision arithmetic, enabling faster inference without sacrificing performance.

Contribution

It proposes a new architecturally-agnostic quantization-aware training method for GNNs that improves accuracy and generalization over existing baselines.

Findings

INT8 models match FP32 performance in most cases

INT4 models achieve up to 26% gains over baselines

up to 4.7x CPU speedup with INT8 inference

Abstract

Graph neural networks (GNNs) have demonstrated strong performance on a wide variety of tasks due to their ability to model non-uniform structured data. Despite their promise, there exists little research exploring methods to make them more efficient at inference time. In this work, we explore the viability of training quantized GNNs, enabling the usage of low precision integer arithmetic during inference. We identify the sources of error that uniquely arise when attempting to quantize GNNs, and propose an architecturally-agnostic method, Degree-Quant, to improve performance over existing quantization-aware training baselines commonly used on other architectures, such as CNNs. We validate our method on six datasets and show, unlike previous attempts, that models generalize to unseen graphs. Models trained with Degree-Quant for INT8 quantization perform as well as FP32 models in most cases; for INT4 models, we obtain up to 26% gains over the baselines. Our work enables up to 4.7x speedups on CPU when using INT8 arithmetic.