Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference

Citations Over TimeTop 10% of 2018 papers

Abstract

The rising popularity of intelligent mobile devices and the daunting computational cost of deep learning-based models call for efficient and accurate on-device inference schemes. We propose a quantization scheme that allows inference to be carried out using integer-only arithmetic, which can be implemented more efficiently than floating point inference on commonly available integer-only hardware. We also co-design a training procedure to preserve end-to-end model accuracy post quantization. As a result, the proposed quantization scheme improves the tradeoff between accuracy and on-device latency. The improvements are significant even on MobileNets, a model family known for run-time efficiency, and are demonstrated in ImageNet classification and COCO detection on popular CPUs.

Related Papers

• → Deep Residual Learning for Image Recognition(2016)216,943 cited
• → Very Deep Convolutional Networks for Large-Scale Image Recognition(2014)75,407 cited
• → MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications(2017)9,892 cited
• → DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients(2016)1,796 cited