Pruning and Quantization Impact on Graph Neural Networks

TL;DR

This paper empirically investigates how pruning and quantization techniques affect the size, accuracy, and efficiency of graph neural networks across various tasks and datasets.

Contribution

It provides a comprehensive analysis of three pruning and three quantization methods on multiple GNN models and datasets, highlighting their effects on model compression and performance.

Findings

Unstructured pruning can reduce model size by 50% with maintained or improved accuracy.

Quantization impacts vary across datasets and tasks, affecting accuracy and inference time.

Different compression techniques offer trade-offs between size reduction and model performance.

Abstract

Graph neural networks (GNNs) are known to operate with high accuracy on learning from graph-structured data, but they suffer from high computational and resource costs. Neural network compression methods are used to reduce the model size while maintaining reasonable accuracy. Two of the common neural network compression techniques include pruning and quantization. In this research, we empirically examine the effects of three pruning methods and three quantization methods on different GNN models, including graph classification tasks, node classification tasks, and link prediction. We conducted all experiments on three graph datasets, including Cora, Proteins, and BBBP. Our findings demonstrate that unstructured fine-grained and global pruning can significantly reduce the model's size(50\%) while maintaining or even improving precision after fine-tuning the pruned model. The evaluation of different quantization methods on GNN shows diverse impacts on accuracy, inference time, and model size across different datasets.