{"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/squeezed-very-deep-convolutional-neural","title":"Squeezed Very Deep Convolutional Neural Networks for Text Classification","arxiv_id":"1901.09821","date":"2019-01-28","proceeding":null,"authors":["Andréa B. Duque","Luã Lázaro J. Santos","David Macêdo","Cleber Zanchettin"],"abstract":"Most of the research in convolutional neural networks has focused on\nincreasing network depth to improve accuracy, resulting in a massive number of\nparameters which restricts the trained network to platforms with memory and\nprocessing constraints. We propose to modify the structure of the Very Deep\nConvolutional Neural Networks (VDCNN) model to fit mobile platforms constraints\nand keep performance. In this paper, we evaluate the impact of Temporal\nDepthwise Separable Convolutions and Global Average Pooling in the network\nparameters, storage size, and latency. The squeezed model (SVDCNN) is between\n10x and 20x smaller, depending on the network depth, maintaining a maximum size\nof 6MB. Regarding accuracy, the network experiences a loss between 0.4% and\n1.3% and obtains lower latencies compared to the baseline model.","url_abs":"http://arxiv.org/abs/1901.09821v1","url_pdf":"http://arxiv.org/pdf/1901.09821v1.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":"squeezed-very-deep-convolutional-neural","repo_url":"https://github.com/lazarotm/SVDCNN","is_official":0,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"classification-1","task_name":"Classification"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"sentiment-analysis","task_name":"Sentiment Analysis"},{"task_slug":"text-classification","task_name":"Text Classification"},{"task_slug":"text-classification-1","task_name":"text-classification"}],"methods":[{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"global-average-pooling","method_name":"Global Average Pooling"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/sentiment-analysis-on-yelp-binary","task":"Sentiment Analysis","dataset":"Yelp Binary classification","model":"SVDCNN","rank_in_archive_order":19,"of":20,"metrics":{"Error":"4.74"},"uses_additional_data":false},{"leaderboard":"/sota/sentiment-analysis-on-yelp-fine-grained","task":"Sentiment Analysis","dataset":"Yelp Fine-grained classification","model":"SVDCNN","rank_in_archive_order":17,"of":17,"metrics":{"Error":"46.80"},"uses_additional_data":false},{"leaderboard":"/sota/text-classification-on-ag-news","task":"Text Classification","dataset":"AG News","model":"SVDCNN","rank_in_archive_order":18,"of":24,"metrics":{"Error":"9.45"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}