{"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/paying-more-attention-to-attention-improving","title":"Paying More Attention to Attention: Improving the Performance of Convolutional Neural Networks via Attention Transfer","arxiv_id":"1612.03928","date":"2016-12-12","proceeding":null,"authors":["Sergey Zagoruyko","Nikos Komodakis"],"abstract":"Attention plays a critical role in human visual experience. Furthermore, it\nhas recently been demonstrated that attention can also play an important role\nin the context of applying artificial neural networks to a variety of tasks\nfrom fields such as computer vision and NLP. In this work we show that, by\nproperly defining attention for convolutional neural networks, we can actually\nuse this type of information in order to significantly improve the performance\nof a student CNN network by forcing it to mimic the attention maps of a\npowerful teacher network. To that end, we propose several novel methods of\ntransferring attention, showing consistent improvement across a variety of\ndatasets and convolutional neural network architectures. Code and models for\nour experiments are available at\nhttps://github.com/szagoruyko/attention-transfer","url_abs":"http://arxiv.org/abs/1612.03928v3","url_pdf":"http://arxiv.org/pdf/1612.03928v3.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":"paying-more-attention-to-attention-improving","repo_url":"https://github.com/szagoruyko/attention-transfer","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"paying-more-attention-to-attention-improving","repo_url":"https://github.com/AlexandrosFerles/NIPS_2019_Reproducibilty_Challenge_Zero-shot_Knowledge_Transfer_via_Adversarial_Belief_Matching","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"paying-more-attention-to-attention-improving","repo_url":"https://github.com/Marior87/Telegram_Bot","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"paying-more-attention-to-attention-improving","repo_url":"https://github.com/bharatsau/Darknet2","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"paying-more-attention-to-attention-improving","repo_url":"https://github.com/chumingqian/Model_Compression_For_YOLOV3-V4","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"paying-more-attention-to-attention-improving","repo_url":"https://github.com/jiahuiiiiii/adapi","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"paying-more-attention-to-attention-improving","repo_url":"https://github.com/yoshitomo-matsubara/torchdistill","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"knowledge-distillation","task_name":"Knowledge Distillation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/knowledge-distillation-on-imagenet","task":"Knowledge Distillation","dataset":"ImageNet","model":"AT (T: ResNet-34 S:ResNet-18)","rank_in_archive_order":51,"of":52,"metrics":{"":"70.70"},"uses_additional_data":false},{"leaderboard":"/sota/knowledge-distillation-on-imagenet","task":"Knowledge Distillation","dataset":"ImageNet","model":"AT  (T: ResNet-34  S:ResNet-18)","rank_in_archive_order":52,"of":52,"metrics":{"CRD training setting":"✓"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1612.03928","atlas_url":"https://app.syntology.ai/?focus=1612.03928","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1612.03928"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/yoshitomo-matsubara/torchdistill","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/AlexandrosFerles/NIPS_2019_Reproducibilty_Challenge_Zero-shot_Knowledge_Transfer_via_Adversarial_Belief_Matching","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/jiahuiiiiii/adapi","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/chumingqian/Model_Compression_For_YOLOV3-V4","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/Marior87/Telegram_Bot","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/bharatsau/Darknet2","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/szagoruyko/attention-transfer","reach":null}],"summary":{"ran_fixture":1},"by_repo_kind":{"official":{"samples":1,"ran":1,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":1,"samples":[{"code_sha256_prefix":"c89b4b7d35ce8ba0","entry":"define_teacher","repo":"szagoruyko/attention-transfer","repo_kind":"official","path":"imagenet.py","file_url":"https://github.com/szagoruyko/attention-transfer/blob/HEAD/imagenet.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"c89b4b7d35ce8ba0"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}