{"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/ultimate-tensorization-compressing","title":"Ultimate tensorization: compressing convolutional and FC layers alike","arxiv_id":"1611.03214","date":"2016-11-10","proceeding":null,"authors":["Timur Garipov","Dmitry Podoprikhin","Alexander Novikov","Dmitry Vetrov"],"abstract":"Convolutional neural networks excel in image recognition tasks, but this\ncomes at the cost of high computational and memory complexity. To tackle this\nproblem, [1] developed a tensor factorization framework to compress\nfully-connected layers. In this paper, we focus on compressing convolutional\nlayers. We show that while the direct application of the tensor framework [1]\nto the 4-dimensional kernel of convolution does compress the layer, we can do\nbetter. We reshape the convolutional kernel into a tensor of higher order and\nfactorize it. We combine the proposed approach with the previous work to\ncompress both convolutional and fully-connected layers of a network and achieve\n80x network compression rate with 1.1% accuracy drop on the CIFAR-10 dataset.","url_abs":"http://arxiv.org/abs/1611.03214v1","url_pdf":"http://arxiv.org/pdf/1611.03214v1.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":"ultimate-tensorization-compressing","repo_url":"https://github.com/timgaripov/TensorNet-TF","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"ultimate-tensorization-compressing","repo_url":"https://github.com/Gyiming/MobileSLAM","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1611.03214","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}