Training with reduced precision of a support vector machine model for text classification

TL;DR

This paper investigates how quantization and reduced precision training of support vector machines affect efficiency and accuracy in multi-class text classification, demonstrating potential hardware benefits.

Contribution

It provides a comparative analysis of SVM training with various reduced precisions on CPU and GPU, highlighting efficiency gains and accuracy impacts.

Findings

Reduced precision decreases training time and memory usage.

Quantization impacts classification accuracy variably.

GPU implementations with half precision show significant efficiency improvements.

Abstract

This paper presents the impact of using quantization on the efficiency of multi-class text classification in the training process of a support vector machine (SVM). This work is focused on comparing the efficiency of SVM model trained using reduced precision with its original form. The main advantage of using quantization is decrease in computation time and in memory footprint on the dedicated hardware platform which supports low precision computation like GPU (16-bit) or FPGA (any bit-width). The paper presents the impact of a precision reduction of the SVM training process on text classification accuracy. The implementation of the CPU was performed using the OpenMP library. Additionally, the results of the implementation of the GPU using double, single and half precision are presented.