{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/shiftcnn-generalized-low-precision","title":"ShiftCNN: Generalized Low-Precision Architecture for Inference of Convolutional Neural Networks","arxiv_id":"1706.02393","date":"2017-06-07","proceeding":null,"authors":["Denis A. Gudovskiy","Luca Rigazio"],"abstract":"In this paper we introduce ShiftCNN, a generalized low-precision architecture\nfor inference of multiplierless convolutional neural networks (CNNs). ShiftCNN\nis based on a power-of-two weight representation and, as a result, performs\nonly shift and addition operations. Furthermore, ShiftCNN substantially reduces\ncomputational cost of convolutional layers by precomputing convolution terms.\nSuch an optimization can be applied to any CNN architecture with a relatively\nsmall codebook of weights and allows to decrease the number of product\noperations by at least two orders of magnitude. The proposed architecture\ntargets custom inference accelerators and can be realized on FPGAs or ASICs.\nExtensive evaluation on ImageNet shows that the state-of-the-art CNNs can be\nconverted without retraining into ShiftCNN with less than 1% drop in accuracy\nwhen the proposed quantization algorithm is employed. RTL simulations,\ntargeting modern FPGAs, show that power consumption of convolutional layers is\nreduced by a factor of 4 compared to conventional 8-bit fixed-point\narchitectures.","url_abs":"http://arxiv.org/abs/1706.02393v1","url_pdf":"http://arxiv.org/pdf/1706.02393v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"shiftcnn-generalized-low-precision","repo_url":"https://github.com/gudovskiy/ShiftCNN","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"caffe2","reach":null}],"tasks":[{"task_slug":"quantization","task_name":"Quantization"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1706.02393","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}