{"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/partial-order-pruning-for-best-speedaccuracy","title":"Partial Order Pruning: for Best Speed/Accuracy Trade-off in Neural Architecture Search","arxiv_id":"1903.03777","date":"2019-03-09","proceeding":"CVPR 2019 6","authors":["Xin Li","Yiming Zhou","Zheng Pan","Jiashi Feng"],"abstract":"Achieving good speed and accuracy trade-off on a target platform is very\nimportant in deploying deep neural networks in real world scenarios. However,\nmost existing automatic architecture search approaches only concentrate on high\nperformance. In this work, we propose an algorithm that can offer better\nspeed/accuracy trade-off of searched networks, which is termed \"Partial Order\nPruning\". It prunes the architecture search space with a partial order\nassumption to automatically search for the architectures with the best speed\nand accuracy trade-off. Our algorithm explicitly takes profile information\nabout the inference speed on the target platform into consideration. With the\nproposed algorithm, we present several Dongfeng (DF) networks that provide high\naccuracy and fast inference speed on various application GPU platforms. By\nfurther searching decoder architectures, our DF-Seg real-time segmentation\nnetworks yield state-of-the-art speed/accuracy trade-off on both the target\nembedded device and the high-end GPU.","url_abs":"http://arxiv.org/abs/1903.03777v2","url_pdf":"http://arxiv.org/pdf/1903.03777v2.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":"partial-order-pruning-for-best-speedaccuracy","repo_url":"https://github.com/lixincn2015/Partial-Order-Pruning","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"partial-order-pruning-for-best-speedaccuracy","repo_url":"https://github.com/zym1119/Partial-Order-Pruning-Demo","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":null,"task_name":"GPU"},{"task_slug":"architecture-search","task_name":"Neural Architecture Search"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"}],"methods":[{"method_slug":"speed","method_name":"SPEED"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/semantic-segmentation-on-bdd100k-val","task":"Semantic Segmentation","dataset":"BDD100K val","model":"DF1-Seg","rank_in_archive_order":15,"of":24,"metrics":{"mIoU":"42.5(82.3fps)"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-bdd100k-val","task":"Semantic Segmentation","dataset":"BDD100K val","model":"DF2-Seg","rank_in_archive_order":16,"of":24,"metrics":{"mIoU":"47.8(53.4fps)"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1903.03777","atlas_url":"https://app.syntology.ai/?focus=1903.03777","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}