{"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/fixynn-efficient-hardware-for-mobile-computer","title":"FixyNN: Efficient Hardware for Mobile Computer Vision via Transfer Learning","arxiv_id":"1902.11128","date":"2019-02-27","proceeding":null,"authors":["Paul N. Whatmough","Chuteng Zhou","Patrick Hansen","Shreyas Kolala Venkataramanaiah","Jae-sun Seo","Matthew Mattina"],"abstract":"The computational demands of computer vision tasks based on state-of-the-art\nConvolutional Neural Network (CNN) image classification far exceed the energy\nbudgets of mobile devices. This paper proposes FixyNN, which consists of a\nfixed-weight feature extractor that generates ubiquitous CNN features, and a\nconventional programmable CNN accelerator which processes a dataset-specific\nCNN. Image classification models for FixyNN are trained end-to-end via transfer\nlearning, with the common feature extractor representing the transfered part,\nand the programmable part being learnt on the target dataset. Experimental\nresults demonstrate FixyNN hardware can achieve very high energy efficiencies\nup to 26.6 TOPS/W ($4.81 \\times$ better than iso-area programmable\naccelerator). Over a suite of six datasets we trained models via transfer\nlearning with an accuracy loss of $<1\\%$ resulting in up to 11.2 TOPS/W -\nnearly $2 \\times$ more efficient than a conventional programmable CNN\naccelerator of the same area.","url_abs":"http://arxiv.org/abs/1902.11128v1","url_pdf":"http://arxiv.org/pdf/1902.11128v1.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":"fixynn-efficient-hardware-for-mobile-computer","repo_url":"https://github.com/ARM-software/DeepFreeze","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"},{"task_slug":"image-classification","task_name":"image-classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}