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.
| Comments: | 14 pages, 12 figures |
| Subjects: | Machine Learning (cs.LG); Machine Learning (stat.ML) |
| Cite as: | arXiv:1712.05877 [cs.LG] |
| (or arXiv:1712.05877v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.1712.05877 arXiv-issued DOI via DataCite |
Submission history
From: Bo Chen [view email]
[v1]
Fri, 15 Dec 2017 23:56:52 UTC (392 KB)